How bilingual kids transfer skills between languages (and how you can help)

Researchers, parents, and teachers often talk about bilingual language development in terms of one language or the other, but not combined. Parents might worry that their child has more vocabulary in English than in Spanish, or that they switch languages halfway through a story or when speaking to different people.

We place a lot of focus on the idea that a bilingual person is essentially “two monolinguals in one,” but experimental and anecdotal evidence increasingly demonstrates that this is simply not true. Bilingual people have unique competencies that should be nurtured and celebrated — not just as far as language goes, but also in more general social and cognitive domains such as attention span, inhibitory control, and cultural awareness.

Today I’m going to talk about linguistic interdependence and how you can help your bilingual child, friend, or self embrace the gift of multilingualism, from the perspective of language acquisition research.

What is linguistic interdependence?

Linguistic interdependence is the idea that certain skills and knowledge can be shared between languages. There are certain abstract features and patterns that underlie human languages, and bilinguals are able to tap into their knowledge of these features in different ways depending on the language they are currently using.

This is similar to but not quite the same as the concept of language transfer, which was popularized in the adult language acquisition literature in the 90s. Cole and Sprouse’s “full access/full transfer” (FA/FT) hypothesis, which proposes that all of the knowledge from one’s native language is available and can be used when they learn an additional language. This can of course be helpful in some ways (namely, when characteristics of the two languages are the same) and detrimental in others.

What distinguishes linguistic interdependence from language transfer is the notion that someone who uses multiple languages draws from a pool of non-language specific knowledge. It also focuses more on what bilinguals stand to gain from their competence in multiple languages, whereas transfer is usually used to explain why language learners make certain mistakes.

Language interdependence is a concept that highlights the interconnected nature of bilinguals’ linguistic skills. In other words, it shows how proficiency in one language can support and enhance the development of the other language.

Linguistic interdependence IRL

Let’s start by looking at reading. Every time I start to learn a new language, I don’t have to completely start over my literacy journey from scratch. Of course, I have to learn new vocabulary and perhaps a new alphabet, but honestly those are only a few of the many skills that children develop in elementary school. Some other critical prerequisites might include:

• How a book works: where is the front, where is the back, how/when do you turn the pages, what information is contained in different sections, etc.
• The structure of a narrative (beginning, middle, end)
• Different types of literature and their features (fiction/non-fiction)
• Inferring the meaning of new vocabulary from context
• Identifying the main ideas of a text

and so on. These are all non-language-specific skills (setting aside some finer features of book layouts), which means that bilingual children can access them no matter which language they are actively using! In other words, building literacy skills in one language can indirectly help the other language.

Looking at my favorite area, phonology, we can see more evidence for linguistic interdependence. Another critical prerequisite for learning to read is a skill called phonological awareness. This is basically the ability to break words up into individual sounds, like how cat can be broken up as /k/ /a/ /t/. Children who have developed phonological awareness in one language will not need to start over again completely in another: they will retain the knowledge that words are made up of individual, smaller sounds, and that alphabets represent those sounds in a written way.

Bilinguals’ knowledge of vocabulary or word structure in one language may also enrich their self-expression in the other language. For example, in American Sign Language, some signs use the first letter of the English word to distinguish semantically related terms from one another, e.g. group, team, and family are the same sign but with a G handshape, a T, and an F, respectively.

These types of similarities between words can highlight the semantic link between words, while also providing a useful mnemonic for ASL-English bilinguals. Bilinguals can draw on their knowledge of prefixes, suffixes, or general word structure in one language to infer the meaning of an unfamiliar word in another language (like how I learned the word “augment” in English from Spanish “aumentar”).

Practical Tips for Parents and Teachers

Compare languages

Ask your child questions that prompt them to think about the similarities and differences between their two (or more) languages. Some examples:

• Which word has more sounds in it: cat or gato? What are they?
• Dimsum in English sounds a lot like dim²sam¹ in Cantonese! Why do you think that is?
• When I ask a question in German, I put the verb at the front of the sentence. What do I do in English? (Or: ask them to translate a sentence and then ask why the same words don’t appear in the same places)
• Why do Spanish and Dutch sound different? Are there any way that they sound the same?

If you are also bilingual, you can model this behavior by simply sharing your observations about the languages you speak.

Watch the same movie in both languages

Pick a movie (if you hate dubbing, then try a cartoon where it won’t be as obvious that the mouth movements aren’t synchronized to the words) and watch it once in Language A, then a few nights later in Language B. Then, discuss the movie with them by asking questions such as:

• What happened in the movie? Can you tell me again in Language A/B?
• Which version did you like more? Why?
• Do you think [important/favorite scene] felt the same when we watched it in Language B? Why?
• Did you think it was easier to understand the movie the second time around? How could that be, if the movie wasn’t in the same language?
• The name in Language A is X, but in Language B it’s Y. Why would they change the name?

Read

Read with them, read to them, read around them, and buy them books. Don’t force them to only read in one language or another; just be glad they are reading!

Encourage them to use all of their linguistic skills

If a child is more capable of explaining something in one language, let them! At the end of the day, language is a tool that allows us to share our complex inner worlds with one another. By allowing them to use all their tools freely, you give your child the best possible chance to express their genuine self.

In an educational setting, provide books or resources in languages that students in your classroom might use outside of school. Allow them to demonstrate their expertise in the language they feel most comfortable using, and consider all of their capabilities when assessing their progress.

Don’t stress about a particular language

On that note, a lot of people get caught up in micromanaging the exact amount of exposure their child will have to each language. Underlying this common concern is the idea that anything other than a perfect 50/50 split will result in an “imperfect” bilingual. The truth is, bilingualism is not a discrete characteristic; it is a spectrum of abilities, knowledge, and comfort. And again, bilinguals are not simply two monolinguals stacked on top of each other in a trench coat.

Just because someone is more comfortable discussing nuclear physics in German and the dinner menu in Spanish doesn’t make them any less of a bilingual. Domain-specific knowledge is perfectly normal. Of course, huge discrepancies in the amount of opportunity someone has to practice can affect their depth of knowledge in one language, but sometimes that is the most natural outcome. Bilingualism is a spectrum, language learning is always good, and imperfections make us human.

Keep a positive attitude

Maintain a positive attitude toward different languages and cultures and encourage your child to use all of their languages in various contexts (playdates, cultural events, etc.). If you’re also multilingual, model your use of all your languages and talk about how it positively impacts your life.

If you are a teacher, offer your students opportunities to share insights from their other language(s) in a classroom setting. Give them the chance to teach you or their peers some words, explain a grammatical structure, or do a report on a book in another language.

Conclusion

Linguistic interdependence can play a significant role in bilingual children’s language development, as skills and knowledge in one language can positively influence the learning and proficiency in the other language. It is time we stopped thinking of bilingualism in terms of two discrete linguistic systems and instead viewed it as an ever-expanding network of competencies. By doing so, we encourage the next generation of global ambassadors to embrace their unique skillset and reveal their authentic selves.

Selected References/Further Reading

De Houwer, A. (2009). An introduction to bilingual development (Vol. 4). Multilingual Matters.

Gawlitzek-Maiwald, I. & Tracy, R. (1996). Bilingual bootstrapping. Linguistics, 34(5), 901-926. https://doi.org/10.1515/ling.1996.34.5.901

Pearson, B. Z. (2008). Raising a bilingual child. Random House Reference.

Tong, X., Kwan, J.L.Y., Xiuli Tong, S., & Deacon, S.H. (2022). How Chinese–English Bilingual Fourth Graders Draw on Syntactic Awareness in Reading Comprehension: Within- and Cross-Language Effects. Read Res Q, 57(2), 409– 429. https://doi.org/10.1002/rrq.400

When Teepee Becomes Peepee

My nibling is producing real, actual words! …mostly. Like most toddlers, the quality of production leaves something to be desired. Teepee, for example, comes out more like “peepee,” and coffee sounds more like “cocky.”

Obviously, the parents are having a great time, and I’m dusting off my background in phonology to analyze every single word they say.

It’s a fairly common stage in language development: children go through phases where they consistently replace the consonants in a word (the “target form”) with one from elsewhere in the word. This practice is called consonant harmony, and it was very intriguing for linguists in the 90s because it rarely happens in adult languages, except when the consonants are right next to each other (e.g. pronouncing input as “imput”). On the other hand, vowel harmony — where all the vowels in a word become the same — is a pretty common process across the world’s languages, even when the target vowels are separated by one or more consonants.

This observation is a Big Deal for linguists who believe that child language never deviates from the possible permutations seen in adult language. One of the key ideas in generative grammar is that, even if children are making lots of mistakes as they acquire their language, the grammatical choices they are making actually correspond to completely valid possibilities in some other language. A kid who says “Want cookie” is not following the rules of English (which requires a subject), but is perfectly fine in Spanish.

So, if a child says “peepee” instead of teepee, and that kind of change doesn’t happen in adult languages, that’s pretty dang interesting.

What’s going well

They have successfully identified and produced lots of important phonetic features: they know that the /p/ and /t/ sounds are both produced without vibrating the vocal cords; they know that /p/ and /t/ should be followed by a small puff of air (= aspiration) when they appear at the beginning of a word; and they know that both of these sounds are the result of fully preventing air from escaping the lungs for a split second (= stop consonants), as opposed to only partially limiting airflow as you would in a sound like /f/.

In fact, the only difference between /p/ and /t/ is that one is produced on the lips, while the other is produced with the tongue behind the teeth.

For a long time, everyone more or less assumed that these supposed speech errors were random, the result of children being unable to perceive or produce the subtle distinctions between sounds of the same category (i.e. stops).

However, as I’ve mentioned previously on this blog, it’s not true that babies can’t hear the differences between sounds. They’re actually really good at it. More counterevidence comes from the “fis” phenomenon, which was a study in the 60s where researchers showed a toy fish to a child who consistently “fis” instead of fish. When they asked him, “Is this a fis?” the child answered no, rejecting the adult’s mispronunciation. This is further evidence that the ability to perceive sounds precedes the ability to produce them.

This probably means that whatever goes awry happens somewhere in the production process: that is, the child has a mental representation of the word teepee in their head, but something gets jumbled up when they try to coordinate all the muscles needed to produce the word.

What’s interesting, though, is that there seem to be fairly consistent rules governing child consonant harmony. The fact that these rules exist is a testament to just how much children already know about the language(s) spoken around them.

Why doesn’t “teepee” end up as “teetee?”

Let’s take a quick dive into linguistic theory. Optimality Theory, proposed by Prince and Smolensky in the early 90s, is used in phonology to explain why certain sound alterations occur in a given context. The idea is that that there are underlying “rules” that everyone has in their heads, ranked in order of importance. Higher-ranked rules cannot be broken, whereas lower-ranked rules can be broken if there is no other choice. The ultimate goal: produce a word that is as close to the representation in your head as possible.

Lots of these rules target not just one sound, but a group of sounds that share a feature (such as place of articulation). When we group the consonants in question according to where they are produced, then we have the following categories.

Name	Place of articulation	Stop consonants
Labials	The lips	p, b, (m)
Coronals	The tip of the tongue	t, d, (n)
Velars	The velar ridge	k, g

An example of a rule that contributes to consonant harmony could be something like “Preserve all labials in the target word.” When a child then cannot stay true to the adult form of the word (for whatever reason), then they might refer back to the hierarchy of rules to decide which consonants they should prioritize.

Once they decide which segments can be changed and which ones have to stay the same, they can transform the word in some way that makes it easier to produce. This is a topic that I looked at for my Master’s thesis, so I can tell you firsthand that lots of children prefer to simply omit the challenging sound(s) altogether. This is why spot might become “pot,” or green might become “geen.”

Another option, of course, is to repeat a sound that is already available to you, especially if they share some phonetic features with the sound you want to (but can’t) produce. This is where consonant harmony may occur.

Since kids show different orders of preference for labials, coronals, and velars at this stage, the segment they tend to preserve can show us which consonant types they find most important (or maybe just easier to pronounce). This, in turn, can give us a hint about the ranking of rules in their head at any given stage. It’s time to analyze some data!

Looking back at “peepee,” it seems like labial consonants (/p, b, m/) outrank coronals (/t, d/), because the /p/ sound is what remains. Based on this information, we can make some predictions about how my nibling would pronounce other words at this stage:

• Potty > “poppy”
• Table > “babu” (/l/ is a whole other can of worms, so just go with me on this)
• Tuba > “buba”
• Boot > “boop” or “boo,” depending on whether deletion or assimilation is preferred at the end of a word

Pesky vowels

Now we have our first set of predictions, but they might need to be adjusted as we get more info. Where do velars (/k, g/) rank in comparison to labials and coronals?

Again, we can determine the order of priority based on how certain words are pronounced. Luckily, said nibling is in the middle of a vocabulary explosion and has given us two pieces of data to work with:

• Doggy > “goggy”
• Digger > “dadu”

Uhh… what gives?! In doggy, the /g/ stays, but in digger, the /d/ stays! There is no preference — it’s all just random after all!

This is the fun of analyzing linguistic data: just when you think you have a working hypothesis, something comes along to throw you off course.

You might be happy (or disappointed, if you were rooting against me) to know that there is a potential explanation for these two data points, and it has to do with the phonetic environment — that is, the other sounds that are occurring around the consonants in question. In our case, I suspect the first vowel is the culprit.

The /o/ in doggy is produced toward the back of the mouth, quite close to the velar ridge. On the other hand, the /i/ in digger (or the /æ/ in “dadu”) is articulated toward the front of the mouth. So, it seems plausible that in words featuring a front vowel like /i/, the coronal sound might be preserved, whereas in words containing a back vowel like /o/, the velar sound wins out. Let’s make some more predictions based on this hypothesis:

• Kitty > “titty” (…it’s what the science dictates, okay?)
• Taco > “kako” (according to American English vowels)
• Gouda > “duda”

Since we only have, like, three words to work with at the moment, it’s hard to say for certain whether the vowels would also have an impact on consonant harmony in words containing a labial consonant. If the vowel has no impact, then we can assume that the labials would win out over velars. If the vowel does have an effect, then we can assume that front vowels would trigger labials, while back vowels would trigger velars.

A few more predictions:

• Piggy > “pibby”
• Copper > “kako”
• Bugle > “bugu”
• Baggy > “babby”

As an aside, it’s very hard to come up with words that a baby might know that also have a (roughly) CVCV structure. So please excuse the ridiculous examples above.

How is this useful?

Sometimes it’s just nice to be reminded that your child is actually supremely intelligent and already pretty adept at this “language” thing. They’ve laid a foundation in their native language(s) and know how it should sound, the implementation is the only problem.

You can also start to appreciate the steps that they take on their unique language journey. Instead of just waking up one day as a fully competent communicator, they make small strides every day that all eventually add up to adult competence. It’s cool to observe the process.

Or maybe you’re just like me and you love finding patterns in everything and solving puzzles. That’s also a perfectly good reason to learn anything. If you’re bored with the monotony of childcare one day, you can always whip out a notebook and start trying to figure out the patterns in their speech errors.

Until next time!

白云苍狗 perception is ever-evolving

白 bái: white

云 yún: cloud

苍 cāng: gray

狗 gǒu: dog

Surprise — it’s not a war story! This idiom comes from a poem by Tang dynasty poet Du Fu.

There was once an aspiring poet named Wang Liyou whose wife left him because he was poor. The town gossip mill started churning and people began to suspect that Wang Liyou had been the one to leave his wife because he was having an affair.

Du Fu, however, could see the truth of the situation and subsequently wrote a poem about Wang Liyou’s troubles. The poem is called 可吹 kě chuī and it contains an iconic line about the ever-changing nature of human perception. Du Fu points out that clouds in the sky can take on many forms, and often just as quickly as we recognize a familiar shape in them, they change form once again. Thus, a fluffy white cloud can just as easily be perceived as a gray dog.

按兵不动 To bide one’s time

按 àn: to press down (e.g. a button); to hold

兵 bīng: troops, soldiers

不 bù: no/not

动 dòng: to move

Lots of Chinese idioms come from war stories, and this is indeed one of them.

During the Spring and Autumn period (770–476 BCE), the kingdom of Jin was plotting against the kingdom of Wei. Zhao Jianzi was planning a military attack against Wei and ordered a government official, Shi Mo, to run ahead to Wei, determine the strength of their forces, and report back within a month.

A month came and went with no word from Shi Mo. In fact, it took six whole months for him to return.

Obviously, Zhao Jianzi was pissed and demanded to know why it had taken Shi Mo so long. Shi Mo told him that times had changed in Wei — their military is now stronger than ever, and the people are feeling very patriotic. Confucius and his disciples are now also wholehearted supporters of Wei (that’s quite a Big Deal™️). Shi Mo concluded that it would be extremely difficult to overthrow Wei at this point in time.

Zhao Jianzi, being the wise warmonger that he was, heeded Shi Mo’s advice. He 按兵不动 — ordered his troops not to move, to wait for a more opportune time to overthrow Wei.

Edited 11.12.22: Added dates and changed a name — it was actually the state of Jin that targeted Wei.

Why can’t we “hear” sounds from another language?

You’ve been learning a language for months, you’ve built up a pretty decent vocabulary, but there are just some words that cause you trouble almost any time you use them. UGH.

Unfortunately and ironically, for my parents, one of these words was part of our address in China. A decade of struggles later, nearly half the time we got in a taxi, my dad would make his best attempt and the following conversation would ensue:

• Dad: “Xiù yáng lù.”
• Driver (confused): “Xiù yáng lù?”
• Dad: “Duì. Xiù yáng lù.”
• Driver: “There’s no such thing as xìu yáng lù.”
• Mom: “Xiù yáng lù.”
• Driver (perplexed): “Eh??????”
• Me: “It’s xìu yán lù.”
• Driver: “Oooooh, okay!”
• Dad: “THAT’S EXACTLY WHAT I SAID.”

Various taxi drivers and I tried to clue them in over the years, but alas — the psychology of language was against them.

Why oh why can’t we hear sounds that native speakers can distinguish easily? It all comes down to our unique neuropsychology.

You see, humans are capable of something called categorical perception, which basically means that we can tell when two things belong to the same category, even if they aren’t exactly the same. Two leaves may not look, feel, or smell identical, but we can examine each one and conclude that they are both, indeed, leaves. Categorical perception allows us (and a selection of other animals including brides and chinchillas) to focus on the features that make them similar, and ignore the differences.

Categorical perception is important for perceiving (and interpreting) the sounds of oral language and the handshapes/locations of signed language. When you say “meow” and I say “meow,” for instance, we will not sound completely identical. One of us might have a higher or lower voice; you might produce the e vowel for just a millisecond longer; or I might raise my intonation at the end to make it a question: “meow?” Although we can hear all of those differences, we also know that they are not important for determining whether the same word (and thus, meaning) is intended. We, as adult English speakers, know that both of our productions of “meow” are equally valid instances of the word that means the sound a cat makes.

On the other hand, there’s no reason that someone who has very little experience with spoken language (i.e., a baby) would assume that every little pronunciation difference doesn’t constitute an entirely new word. If they hear me say “Meow!” and “Meow?” with different intonations, how do they know those aren’t two completely different words?

We are born with the ability to perceive every sound used by oral language, and to distinguish even between sounds that are acoustically very similar. A famous study from the 70s showed that 4-month-old infants exposed only to English can still perceive the difference between three Korean consonants which English-speaking adults usually cannot distinguish. However, by 8 months old, they lose that ability. This is a result of certain neurological and psychological processes that help us focus in on the special features of the language(s) that are used around us — the development of categorical perception.

Categorical perception allows us to draw boundaries along a continuum to establish discrete categories. Vowels are a good example of such a continuum in language. They are essentially produced by opening your mouth and vibrating your vocal cords, using the tensing of your tongue and lips to alter the sound that comes out. The difference between /i/ and /e/ depends only on which part of the tongue is tensed; this fact means that you can switch back and forth between the two sounds in a single breath, in one continuous sound. In fact, all the vowels in oral language can be produced in a single breath, simply by varying the position of your tongue and lips. The same cannot be said of consonants.

I’ll pause here to let you test all that out.

Babies gradually learn to perceive vowels and consonants categorically, meaning that they learn to tolerate variations and figure out where to establish a barrier between sounds like /i/ and /e/. Different languages might have slightly different barriers, even between similar-sounding vowels. Other languages might distinguish sounds that don’t exist in English at all, like a vowel between /i/ and /u/. As a baby, your brain learns to pay attention to acoustic signals (sound waves, frequencies, etc.) that are important for determining whether a sound is /e/ or /i/ — in other words, to take something continuous like an acoustic signal and assign it to discrete categories.

Returning to my parents’ Mandarin conundrum, the problem was as follows: Mandarin has a vowel whose acoustic features are a mixture of the English /a/ in “father” and /e/ as in “egg.” I can hear the difference clear as day, but unfortunately, my parents’ brains long ago decided that they would perceive vowels only according to the boundaries that are relevant for English. The acoustic information that distinguishes this vowel in Mandarin is misinterpreted as /a/ because their brain does not have a better category to assign it to.

As a non-linguistic example, imagine you have never seen a butterfly (or any other flying insect) before. You are tasked with sorting pictures of bugs and birds into their respective categories. It’s pretty straightforward until you encounter the butterfly, and then suddenly it’s not. On one hand, this creature has more than 2 legs, just like the other bugs you’ve seen thus far. On the other hand, it also has wings — that isn’t very bug-like, in your experience. You absolutely have to sort these cards, and you cannot create a new category. What do you do?

(It has just occurred to me that this is the exact problem faced by biologists when they first discovered the platypus, which is an egg-laying mammal, so there’s another example for your records.)

When you learn a new language as an adult, you don’t start from scratch. You approach it with years of experience using, interpreting, and categorizing the components of another language. Unfortunately, this can make for some pretty strong biases. Even if you consciously know that this is not your native language, your subconscious is playing a completely different game, using the tools it has honed for the past however-many years of your life. This can make it very different for you to perceive and produce sounds that do not exist in your native language.

Some studies have shown that explicit training on the phonetic differences between non-native sounds can help adult learners perceive and produce them, but the results tend to vary across language pairs, sounds, learner types, and individuals.

Perhaps one day we’ll be able to tap into the mechanisms that babies use to master their native language(s). Until then, though, try to comfort yourself with the fact that the same neurological processes that make you a poor Mandarin speaker also made you an excellent English speaker — or vice versa.

Babies in Space

Taking a quick break from my compling musings to spew some information about prepositions, spatial cognition, and infant development at you all. Enjoy!

As adults, the way that we carve up the physical world around us seems pretty straightforward: objects can be close to us or far away; they can be in front of or behind us; they can sit on top of or underneath one another. We as English speakers have established that certain types of relationships between ourselves and the physical world are important to share with one another, and we express those relationships linguistically using prepositions like in, on, next to, under, behind, etc.

What you may have noticed, though, is that we don’t tend to focus on all the details of those relationships. For example, a lid on a jar is not quite the same thing as a book on a table: lids tend to interlock with the things they are closing, whereas a book rests freely on the table, with no insertion of their respective surfaces. There are other features that differ between the two scenarios: you might notice that a jar tends to be round and cylindrical, while a book and a table are both long and flat. A jar can contain other items, while a table (usually) cannot. In fact, one of the only things they do have in common is that one item sits on top of another. Nevertheless, as English speakers we have mutually concluded that the most salient characteristic of both these relationships is the A-on-B setup, and thus we use the same preposition (“on”) to talk about both.

A book on a table

A lid on a jar

This leads me to my first point: not all languages assign the same types of preposition (or verb, adjective, or postposition) to the same types of physical relationships. If you’ve ever studied a foreign language, you may have noticed this when trying to translate spatial words from your native language to your target language – sometimes native speakers correct your use of a word that seems perfectly acceptable to you. This is because the way we talk about space is somewhat arbitrary; we are trying to carve up reality in such a way that makes sense to us and other members of our speech community, highlighting the details that we find important and ignoring the ones that aren’t relevant or useful. Thus, some languages might make distinctions where you do not, or they might have completely different ways of talking about space than you are used to.

For example, Korean, which would indeed use two different expressions to talk about the lid on the jar versus the book on the table, doesn’t necessarily care about containment (English “in”) or surface contact (English “on”), but rather tight fit (like a lid on a jar, a coin in a slot, or interlaced fingers) and loose fit (like a book on a table, an apple in a backpack, or a pencil in a drawer).

On the other hand, languages such as German and Dutch have the same basic concept of on, but further distinguish between vertical and horizontal contact, so that the picture on the wall (surfaces are in contact vertically) and the picture on the table (surfaces in contact horizontally) use two different words (an and auf, respectively). So in these languages, the orientation of contact is also important. In other languages, neither the orientation nor the contact is important: in Mandarin, the same word (shàng 上) expresses both above and on, regardless of whether the two surfaces are physically touching.

Suffice it to say, there are a lot of tiny details that can add up to salient differences in the way that we talk about space, and make it really hard to adjust to a second language. It’s actually pretty remarkable that we are able to do so without even thinking when we use our native language!

Thus, we come to my second point: how on earth do babies figure it all out?

Linguists and cognitive psychologists have long been interested in the way that infants learn to 1) represent spatial relationships conceptually, and 2) map those concepts onto language. The way that we address the development of cognition (= general thinking abilities) and language very much depends on the theoretical background you subscribe to.

Which Came First, the Language or the Thought?

Most linguists would hiss and retreat if you were to call linguistics a subdiscipline of cognitive science, but in reality the two fields tend to overlap pretty frequently, especially when it comes to language acquisition. After all, language doesn’t exist in a vacuum; we use language to share our thoughts and ideas with one another, and of course those thoughts and ideas are the result of various cognitive processes like abstract thinking, intention reading, and visuospatial representation. When we study infants’ and toddlers’ linguistic development, oftentimes we are faced with a “Chicken vs. Egg” scenario, wherein we have to decide whether the linguistic milestones coincide with, precede, or follow the relevant cognitive concepts they represent.

Some people, like Jill de Villiers, have argued that certain language structures actually trigger cognitive development in relevant areas. In other words, the language provides a way to conceptualize and, thus, think about things like false belief or space. If this were the case for spatial language, we would expect that children do not demonstrate any sort of knowledge about space until they start to understand words that express that knowledge. I suppose it would depend on how you test that knowledge, but given that newborns have demonstrated basic understandings of physics and gravity, I am inclined to believe that they do have some concept of space even before language sets in. So, although I am a fan of Jill’s work with regard to theory of mind development, I don’t think we could extend the theory to spatial language.

Alternatively, we could argue that the concepts exist independently of the language, which of course is how we are able to understand the way that another community talks about space without actually speaking the language in question. What do children have to do, then, to link them together? And, is there any rhyme or reason as to how spatial language develops?

In one of the first studies on this topic, Johnston and Slobin (1979) looked at the production of spatial words in English, Italian, Turkish, and Serbo-Croatian, and found that certain concepts tended to emerge at the same time in all four languages, even if those languages do not distinguish between the concepts in the same way. In the first stage of acquisition, the children in their study produced words for the concepts in, on, under, and beside. In Stage Two, they added words for between and behind/in front of, specifically when talking about objects that had an inherent front or back (such as people or houses, but not e.g. a cardboard box). In the final stage, the children were able to talk about the concepts of behind/in front of for all objects, including those that did not have a static front or back. Johnston and Slobin concluded from these findings that the non-linguistic development of the relevant spatial concepts triggers children to look for ways to express them with language.

Melissa Bowerman and her colleagues agreed that infants and toddlers have preconceived ideas about space and location that exist independent of their linguistic knowledge. They support this view with evidence from child speech, namely the fact that children tend to overgeneralize or misuse spatial language early in the acquisition process (e.g. saying “book up” for the book is on the table). However, they have taken this a step further by demonstrating that children as young as 18 months old understand basic spatial language, including the details that are relevant to how their target language expresses ideas about space.

Bowerman and her co-author, Soonja Choi, ran a looking preference task with 14- and 18-month-old infants acquiring either English or Korean. Looking preference paradigms are a common way to test preverbal babies to see how much language they understand, and it works like this: the baby sits with their caregiver in a room, with two screens in front of them. On one screen, there is a video of e.g. a pen being put into a bag, while on the other screen, there is a video of e.g. a book being placed on a table. The baby then hears a sentence like “Put it on there!” The independent variable is how long the baby looks at each screen, because if they understand the sentence, they should look longer at the screen with a scene that matches the language they heard. In control trials, both videos would be valid matches to the sentence, so the baby should not show a preference for either screen.

Using this setup, Bowerman and Choi found that by 18 months old, Korean and English speaking babies were sensitive to the way that their specific language talks about spatial relationships. Korean infants distinguished between loose and tight-fitting relationships, while English babies distinguished between containment (in) and support (on). Furthermore, in trials where the target word indicated a different relationship between the objects in English than Korean (e.g. both objects were on in English, but one was loose and one was tight in Korean), children adhered to language-specific preferences when selecting a video to look at. This suggests that they are not simply mapping the language they hear onto universally-available concepts of space; they are acutely aware of the potential similarities between notions that are expressed the same way linguistically. In other words, they use their amazing pattern-finding abilities to determine what is conceptually similar between a book on a table and a lid on a jar, and they hone in on these differences very early in the acquisition process.

Basic spatial expressions are actually among some of the earliest produced words in many languages, although they are typically not used in a completely adultlike manner. As I said previously, small children may go through a phase of over- or undergeneralizing certain spatial terms while they figure out exactly which concept a given word maps onto. They might also use them a bit more liberally than adults, in some cases, like the infamous “Up!” which is actually more of a verb (denoting the action of being picked up) than a preposition. All of these phases are just stepping stones on their journey toward mastery of their language(s), but each one can actually provide incredible insight as to the state of their linguistic, cognitive, and social development. Next time you talk to a toddler, if you’re so inclined, you can try giving them some basic instructions like “Show me the pen on the table” and see what they do in response.

How could your virtual assistant “spell” in Chinese? (Part 1/2)

** For relevant background on Chinese orthography, read my last post. **

As you may or may not know, in my spare/procrastinating time I like to read about machine learning and computational linguistics. As I do so, it’s made me more and more curious about how certain things in our daily lives work… like virtual assistants. Back when I first started taking a computational linguistics class during my Master’s program, I drafted a whole blog post about how much it taught me to appreciate the mechanisms behind virtual assistants. That appreciation does indeed continue to this day.

A few weeks ago I was reminiscing about a particular Mandarin teacher I had in middle school. Whenever we asked her how to write something, instead of writing it on the board, she would verbally dictate it to us using strokes, radicals, or other characters. It drove us crazy as teen language learners but of course, in hindsight, it really helped me understand the way that characters are constructed.

Anyway, that sparked an interesting idea: namely, what if I were able to ask the lady in my phone for help when I forgot a character nowadays? That would be pretty neat! I started wondering how we could actually make that happen from a linguistic perspective. That, of course, is precisely what I wanted to talk to you about today.

As far as I’m aware, this isn’t an actual function that’s available on any of the major current virtual assistants (although I’m not as familiar with the current offerings in Mainland China). So of course, this is all just a thought exercise – the best kind of exercise, IMO.

What goes into speech recognition

Natural language understanding (NLU) is a huge and rapidly-expanding field that I certainly can’t condense into a single blog post. But just for some appreciation, let’s think for a moment about how computers have to go about understanding spoken human language.

At its core, oral language is just a bunch of acoustic signals organized in a certain way. There are patterns and rules that govern it: for a simple example, the two sounds /p/ and /f/ never appear together at the beginning of a word in English. At the sentence level, there are also certain combinations of words that just never appear together – like the happily. And in other cases, there are groups of words that appear together quite often, like the and woman, or language and acquisition (okay, that one might be a stretch, but it is certainly the case in my sphere of influence). Discovering many of these rules and patterns is a simple matter of statistics and probability: babies amass a huge amount of firsthand data throughout the first few years of life and use their predictive abilities to determine what is and is not a possible word/sentence in their language(s) before they can even speak for themselves.

Computers, as you may know, are even better at math and statistics than babies… or at least, they’re faster. That means that we can teach them to recognize speech using statistical learning, just like babies! All they need is a bunch of data (and maybe some rules).

When it comes to creating a virtual assistant, the basic task is to teach the computer to link a particular sequence of sounds to an action. The tricky thing is that it also has to tolerate natural variation in the speech signal (because every voice is different and can be impacted by environmental factors like the position of the body), syntax (because there are many ways to say the same thing), and word choice. At the same time, the model you’re using shouldn’t be so tolerant that anything can trigger a response. This is the basic task that computational linguists working on NLU need to accomplish.

Generally, an ideally tolerant NLU system can be accomplished with a whole lot of data and machine learning algorithms. Since I’m first and foremost a linguist, the details I’m interested in have more to do with the way the training data can be prepared, understood, and validated, rather than the actual algorithms being used.

Suffice it to say that you need a lot of data (also known as “training data,” which is what you feed into the model), in the form of annotated audio clips to help the algorithm learn how to segment speech, interpret variation, and translate all of that into actionable requests. For a more concrete example of what that means, let’s look back at my original question and consider some potential problems it could present for NLU and, by extension, a virtual assistant.

Setting up the Answer

Before your favorite virtual assistant can respond to a request for information (like how to write a character), it of course needs to be able to access the answer. In this case, that means you need to compile a database of individual characters, their components, and rules for breaking them down. If you need a refresher on the ways that Chinese characters are written, check out my last post.

Learning Radicals

As we know, there are a finite number of radicals (somewhere around 200) used in Chinese characters, many of which evolved from pictographs thousands of years ago. Some radicals have different appearances depending on their position in the character: for example, 日 “sun” may appear more narrow when it makes up one side of a character, as in 明, or it can be short and stout when it appears on the top, as in 早. It might also be smaller or bigger depending on how much real estate it takes up compared to other parts of the character, since all characters should fit into a s square box. If we want the computer to be able to recognize all the radicals, then the different possible forms of each radical should be included in the training dataset. That means you need several instances of each radical so that the algorithm has an opportunity to note its appearance in multiple characters.

Learning Strokes

There are also a finite number of strokes, each with their own name, along with rules for the order in which you combine them. This means that the virtual assistant has to know the names of strokes as well as the proper ways to list them when describing a given character. That isn’t necessarily straightforward because the rules that govern stroke order tend to depend primarily on visuospatial characteristics. In other words, it isn’t an absolute rule that all horizontal strokes have to precede vertical strokes: a vertical stroke might come first if it is to the left of the horizontal one, or some such thing. Certain rules must give way to others, depending on the particular character in question. That feels quite tricky to teach a computer.

Two different learning mechanisms come to mind, depending on a variety of factors. The first option is to teach the virtual assistant these rules, including a rough hierarchy of how they are applied (i.e. when one trumps the other), and then have it dissect a set of characters in a supervised learning paradigm. The independent variables would be the list of rules, along with a set of characters decomposed into their individual strokes (that’s a separate task for another program – let’s just assume it’s already been done). The dependent variable is the order of strokes for a given character. Humans would then need to verify the output to check whether the appropriate stroke order has been proposed for each character. This might be less time-consuming to implement, since only a handful of rules need to be fed into the model, but would require more post-hoc verification. It would also require another program that first dissects characters into an unordered list of strokes.

The second option would essentially require the model to deduce stroke order given a data set consisting of characters and their corresponding stroke orders. This would require more data preparation initially: someone would need to determine the appropriate number of characters to include in the training data (to avoid under- or overfitting the model) and ensure that these characters accurately represent the types of possible layouts in Chinese characters. That could require taking the time to do a visual analysis of the types of possible structures and their frequency. Then they would need to make sure that the set of characters shown to the model as training data is proportional.

In this case, the independent variable would be a complete character, and the dependent variable would be the way the model breaks it down into its stroke, radical, and/or smaller character components. Training data would explicitly provide the computer with examples of how characters can be segmented (e.g. 饿 = [饣, 我]).

Once you have a way to break down characters, you need another database of that links the written forms with their pronunciations. To account for things like interspeaker variation, you might even have multiple speakers of different demographics (old/young, different genders, different geographic locations, etc) saying the same thing. Diverse data helps the model ignore the red herrings (like tone of voice) and focus on what’s really important. Again, assuming you have a virtual assistant that already understands Mandarin, the capacity to segment speech and identify words should already be present, and just needs to be tweaked a bit for our specific purposes.

Understanding the Words

Most words in Modern Mandarin are polysyllabic, combining the meaning of two or more single-syllable morphemes (e.g. 电 “electricity” + 脑 “brain” → 电脑 “computer”). There are also a ton of homophones: a single syllable can actually map onto different morphemes and, therefore, characters. For instance, shì can mean “to be”  是, “thing” 事, “person” 士, or “generation” 世. If this seems unnecessarily confusing, think about bear and bare in English and how easy it is for native speakers to figure out the intended meaning in an actual sentence. Lots of languages deal with homophony to some degree or another. Just like in English, in Chinese, we can clarify which particular meaning of a homophone is meant by using it in a word, sentence, or phrase in which it frequently appears, or where only one alternative makes sense. Going back to the English example, this would be like saying “How do you spell bear, as in ‘I cannot bear it’?”

When confronted with a particular phonetic (= sound) form that could map onto multiple characters, the virtual assistant can do the exact same thing. It can prompt the user according to the most likely target, or else ask the user for clarification.

If the assistant guesses the context (e.g. “Do you mean shì as in shì rén [‘scholar’]?”), then the user simply has to answer yes or no. But how would the assistant decide which one to guess first?

Once again, we can use statistics and probability! Looking back at the list of possible meanings of shì, you might have an intuition that some are more frequently used than others. The meaning “to be,” for instance, probably comes up a lot more often than “generation.” This information could help the virtual assistant determine which meaning to target first: it can ask about the more frequently occurring words first, and then move down the list of possibilities in order of likelihood.

Astute readers may have already noticed that this presents another psycholinguistic issue: more frequently-used characters should also more likely to be remembered, therefore users are presumably less likely to need help writing them by hand. Is the solution then to start with the least frequent character? For some homophones, starting with one or the other might not make a huge difference, if there are only a couple possibilities. However, for a syllable like shì, there are probably 2 dozen possible characters, some of them very niche or archaic. If you instruct the virtual assistant to always start with the most (in)frequent possibility, it might take several minutes of dialogue to arrive at the real target – not a very user-friendly experience.

Perhaps more effective would be simply allowing the user to specify the target character right off the bat, or with prompting if necessary. Most Mandarin speakers would do this anyway when room for confusion exists (for what it’s worth, I would never just ask someone “How do you write shì?” with no other context).

If the virtual assistant either prompts the user to provide context or the user does so in the initial query, then this issue can already be solved by whatever mechanisms exist for handling homophones in other contexts. Like I said, Mandarin has a whole lot of homophones, but in most cases the intended meaning can be discerned easily from context and world knowledge. If the virtual assistant already supports Mandarin as a language, then mechanisms to handle homophones would already be in place. At a basic level, n-gram probabilities (i.e. the likelihood of a certain word appearing in a certain context), the same algorithm that is used for predictive text, could process the user’s request to determine which is the target syllable: namely, the one that most frequently appears in combination with the words around it.

These are some of the word-level considerations that would need to go into a functional character “speller” for virtual assistants. There is, however, yet another level that we have to consider when programming our robot: the syntactic and semantic level. Namely, how does the virtual assistant interpret different word orders, vocabulary choices, and syntactic dependencies?

Tune in next week to find out!

How do you “spell” in Chinese?

It’s no secret that one of the most attractive and intimidating characteristics of modern Chinese languages is the writing system. Characters are, of course, beautiful, which is why so many Westerners love to get things like “chicken soup” tattooed on their biceps. From a language learning perspective, however, Chinese orthography can be incredibly daunting: there are thousands of characters, although only a fraction of those may be required for most communicative functions.

鸡汤

-Some guy at the gym’s arm

Given these facts, it is no wonder that teaching L2 learners to read and write characters is also a hot topic in CFL (Chinese as a Foreign Language) literature. In fact, that’s what I mostly focused on in my CFL classes. I spent a lot of time researching about the best ways to break down the structure of Chinese. The good news is: characters are not completely random, abstract pictograms! They are actually composed of various subcomponents that you can learn to recognize. The bad news is that there are still more components to learn than there are, say, letters in the English alphabet.

But don’t let that stop you! Let’s talk about the building blocks of Chinese characters, and how one goes about using them to dictate characters to one another.

Components of Chinese Characters

Strokes

Starting from the most basic level, characters are composed of individual strokes which can be divided into eight types (the character 永 yong encompasses all of them). They are called strokes because they can be drawn with a single stroke of a pen/brush. Most strokes are straight or nearly-straight lines (except for 点 dian, which is basically a dot), and each one has its own name. Particular combinations of strokes that tend to frequently co-occur may be referred to as a single word that combines the two names (e.g. 竖 shu “a straight vertical line” + 提 ti “a diagonal line drawn from left to right” = 竖提 shu ti)

There is a long-standing tradition of adhering to specific rules dictating the order in which these strokes are added to the page. Nowadays, stroke order rules underlie some of the mechanisms used to enter Chinese characters using a standard keyboard. Some examples of stroke order rules include things like:

1. Top before bottom
2. Left before right
3. Horizontal before vertical
4. Inside before outside

Radicals

Another important component of characters is radicals. Radicals can be standalone characters or a combination of strokes that appear frequently together and denote a fixed meaning. Individual strokes typically do not entail a fixed meaning, but radicals do. For example, the “grass” radical 艹 is used in the names of plants 草, flowers 花, and tea 茶. There is no inherent pronunciation associated with radicals; they only indicate a character’s meaning.

Radicals are a closed class of about 200 members, although some are certainly more common than others. They are sometimes derived from standalone characters (e.g. the “sun” radical is just the character for sun 日) and can therefore appear on their own, but not always. Radicals often appear as components of phonosemantic characters, which leads me to my next point…

Other Characters

That’s right – characters can be recursively embedded! Love that for us.

So, if you’re reading a Chinese text and you encounter a new character, you might be able to break it down stroke-by-stroke, and maybe you can assign it a vague meaning based on its radical. That still doesn’t help you know how to pronounce it, though.

That’s where other characters come in.

An estimated 85% of characters in Modern Standard Mandarin are considered phonosemantic characters, which means that they consist of two parts: one part indicates how it might be pronounced (“phono-“), while the other part indicates the general meaning (“semantic”). Since neither radicals nor strokes have any associated pronunciation, the phonetic component is conveyed using other characters.

For example, the word for “grass” 草 cao is actually composed of a radical 艹 and another character: 早 zao “morning.” Grass and mornings don’t really have much to do with one another semantically, but the pronunciation of zao is pretty close to that of cao.

(To add another layer of complexity, 早 can actually be further broken down into a radical and another character: 日 “sun” and 十 shi “ten.” Fun!)

Explaining how to write a character

Even native Chinese speakers may occasionally forget a character, especially in this day and age of autocorrect and speech-to-text. And of course, learners may often need to ask native speakers to explain how a character is written. In both cases, one person could help by verbally breaking down the target character into individual strokes, radicals, or smaller characters. Let’s look at an example of how that would go.

A: 张三，你知不知道草字怎么写？

B: 我告诉你。上面是草字头，下面是早上的早。

A: Zhangsan, do you know how to write the character cao?

B: I will tell you. On top is the grass radical, and on the bottom is zao as in “morning.”

In this example, the target character is phonosemantic and contains a very common subcomponent, so it is very feasible that Speaker A would not need any more instruction beyond “zao as in ‘morning.'” If they did, of course, then Speaker B could further elaborate:

A: 那么，早怎么写呢？

B: 上面是日，下面是十。

A: 你说哪一个shi? 有很多啊！

B: 中间写横、竖的十。

A: Well, how do you write zao?

B: On top is “sun,” on the bottom is shi.

A: Which shi? There are many!

B: The shi that is written with heng and shu.

Astute readers may have noticed that these strategies still leave room for misinterpretation. You have to have at least a vague idea of how a character is composed and what the end result should look like in order to apply instructions like the ones above. To be honest, there’s no easy way around that fact. But the same could be said about English spelling — you have to know the building blocks (letters) for verbal dictation to be helpful. It just so happens that there are more components to master in a character-based language.

Why am I telling you all this? Well, first and foremost — it’s my blog and I can ramble if I want to. But also, I recently was toying around with a little thought experiment that I wanted to share. This is all necessary preamble. So stay tuned for the next installment, where I brainstorm how this information could be useful for voice assistants looking to beef up their Chinese language support.

9 Reasons Your Child Should Participate in Language Development Research

One day your child comes home from school or camp with yet another a stack of papers to be signed and returned at your earliest convenience. At the top of a sheet stamped with the local university’s logo, you see the heading: “Consent to participate in a research study.” Your child has been invited to participate in a linguistic experiment. You are instantly wary; perhaps you have watched a documentary about the Stanford Prison Experiment, read an article about MK-Ultra, or followed any of the numerous scandals relating to personal data harvesting on social media platforms. Let’s be honest, there are plenty of reasons to mistrust scientists, especially if you or your child are a member of a marginalized community.

That being said, in this day and age, language acquisition experiments are really nothing to be scared of. There are no medical procedures involved, no psychological warfare, and no abuses of data (as far as I’m aware). On the contrary, it’s often pretty fun for the children, not to mention extremely helpful for people whose careers depend on successful research projects. So, I thought that I would put some information out there to attempt to quell the general public’s fear of the word “experiment” and help caregivers make a more informed decision about when and how to allow their child to participate in language acquisition research.

1. Every child has a something unique to contribute

No two children have exactly the same language experiences, and that creates a lot of potential for variation in the way that they acquire and use their target language(s). While some differences — such as mono- versus multilingualism — are more obvious, there are plenty more subtle variations in the ways that children experience language. For instance, the way that adults speak to children (i.e. the words they use, common sentence structures, and even the rate of speech) varies wildly both across and within cultures. In some families, caregivers may engage in a constant dialogue with their children from Day 1, even if the conversation is mostly one-sided during the early years. In other families, talking to an infant who is incapable of responding might be seen as strange. Even birth order can have an effect on the way that children experience language — older siblings are often a significant source of language input for younger children, which means that they provide a lot of primary linguistic data for their younger siblings to work with. In order to obtain a well-rounded view of the ways that language develops and the limits of its variation, we have to take all of these factors into account. Your child’s unique background can contribute to that knowledge in a huge way!

2. They might learn something about the language(s) they speak

One of my favorite things to do with older children is to ask them what they think I was trying to find out after an experiment is completed. They’re almost never correct, but it does prompt them to reflect on the language that they used throughout the course of the activity, and in doing so they often make pretty astute observations. Even when they struggle to understand or produce a particular structure, the repeated exposure throughout the experiment allows them to grapple with the way that a specific meaning is mapped onto a linguistic form. It’s also not completely unusual for children’s performance to improve during the course of an experiment — there is definitely an educational function that should be more systematically studied.

3. All materials are safe and appropriate for use with children

All researchers working with human participants have to submit detailed proposals to an ethics approval board before they are allowed to collect data. For vulnerable populations such as children, the vetting process is even more extensive. All of the experimental materials (meaning the pictures that participants will see, the sentences they will hear, and anything else the child will be exposed to during an experiment) must be submitted and approved by the ethics board of the university and/or country, and these organizations don’t just give the green light to anybody. In the US and Europe (at least the countries I’ve worked in), anyone who will be in direct contact with children must undergo the same screening as public school staff, which typically involves fingerprinting and a background check. Institutions put a lot of effort into making sure that research studies are conducted ethically, and those who violate ethics policies can be subject to extensive disciplinary measures, including legal action.

4. Your data is protected

When your child participates in an experiment, some personal information might be collected in order to be used in the analysis later on, in addition to any of the results from the experiment itself. All of this information must be stored in a manner compliant with data protection laws (i.e. the GDPR in EU countries) to ensure that it is not misused. In practical terms, this means that information like what language(s) your child speaks, their age, and other relevant demographics might be collected and stored on a server at the university, but they are not linked to your child’s name. Instead, each participant in an experiment is assigned a random identifier, like a sequence of numbers, a pseudonym, or some other code. The only document that contains any real name is usually the consent form (since you have to sign it), which is stored as a physical copy in a file cabinet somewhere, or else on a separate, encrypted hard drive. Institutions have strict guidelines about how experimental data can be stored, and the specifics should be outlined on the information sheet that you receive before consenting.

Researchers also cannot use your data for any purpose other than what they told you it would be used for: for example, if the consent form says that the data will be used for publications relating to XYZ Research Project, they can’t then give it to a colleague to use for ABC Project, even if that colleague is in the same department. If it says that the data will be used in scientific publications, then they can’t post excerpts on social media to show their friends how cool their job is. And of course, they can’t sell it to Big Tech to supplement the egregiously low wages that the university pays them. Furthermore, you are allowed to withdraw consent at any point before or after the experiment, in which case the researchers are legally required to promptly delete all your data. All you have to do is email the primary investigator (PI), whose contact information is listed on the information sheet.

5. You might earn some money

Lots of labs have a budget for reimbursing participants, even if they do not complete the entire study. Of course, we aren’t allowed to give children actual money, but we will often give them a small toy or some stickers to thank them for their time. Alternatively, caregivers might receive monetary compensation or a gift card.

6. You can do many of the studies online

Thanks to the pandemic, online studies are more common than ever before. There are numerous websites where you can sign up to participate in developmental research from the comfort of your own home, such as MIT Lookit. That’s a great way to keep the kids occupied for a few minutes during school vacations.

7. It’s fun!

Acquisition researchers spend a lot of time making their experiments as fun and engaging as possible — after all, you aren’t going to get good data if the kids are just trying to finish the activity as quickly as possible. Most experiments nowadays are basically games with an underlying purpose. I always find that when I collect data at schools or summer camps, only a handful of families initially consent before I start coming in regularly. Inevitably, the number of willing participants increases exponentially as I spend more time with the kids, through no conscious effort on my part. Once the children discover that the experiment is basically just a language game, they tell their friends about it, and suddenly half the class is going home and begging their parents to let them participate. I’ve had plenty of kids ask me if they can play the games again (and occasionally I oblige, if I have time). They genuinely enjoy figuring out the “word puzzles” and winning the games that we have in store for them (spoiler alert: the games are rigged so they always win).

8. Findings from language development studies can have huge implications for educational materials

Lots of findings from acquisition research can be used to improve the language and materials that are used in education. Have you ever wondered whether your child really understands the explanations in their schoolbooks? Or how children learn to label abstract grammatical categories like “noun” or “adjective”? Linguistics research can help answer that question and indicate better ways to formulate explanations and teach lessons on grammar, reading, and writing. There are entire conferences dedicated to this exact topic, such as LiDi.

9. Findings from language development studies can have huge implications for diagnosis of language disorders

Did you know that multilingual children are at risk of being incorrectly diagnosed with a language delay? This is because most of the tests and screening materials for language delays are based on the “normal” developmental trajectory of white, upper-middle class, monolingual children. For a long time, it was assumed that bilingual children had weaker linguistic knowledge than same-age monolinguals in each of their respective languages. This was actually because researchers had only accounted for their knowledge of one language, not both. When you consider the words and structures a bilingual child knows in both their languages, oftentimes they come out ahead of their monolingual peers. This trend of misdiagnosis is a big problem for lots of reasons: first, it means that many (misinformed) pediatricians and teachers continue to encourage parents not to teach their child their native language because they believe that it is somehow damaging to their development. Second, it means that resources are misallocated to help children who don’t actually need any help! On the other hand, without adequate understanding of how multilingual children may differ from monolinguals, we might also miss out on some other signs of disordered language that don’t typically present (or present differently) in one population or the other.

Suffice it to say that if caregivers hadn’t consented to let their bilingual children participate in research, we might never have made that critical discovery about bilingual development. If only certain demographics of children are participating in linguistic research (like monolinguals, only children, or professors’ kids), we aren’t obtaining a sample that truly represents the population as a whole, and all of the conclusions we draw might be biased against other demographics.

Bonus: You might help someone get a degree

It may seem obvious, but it is worth noting that lots of language acquisition research is conducted by students, for whom completion of the project is a requisite for their degree. That means if they don’t have participants, they don’t get a degree (or they have to go back to the drawing board and waste potentially years of work). Finding willing participants is one of the hardest parts of conducting acquisition research, and it has only gotten harder in these post-pandemic times. If you have the time and resources, and you think your child would enjoy it, you might consider helping out your friendly neighborhood acquisition researcher.

Proposing the Research

Hello world! Long time no see. You might not be surprised to know that this whole “getting a PhD” thing has been relatively time-consuming, and has left me with minimal energy reserves for frivolous blogging.

Nevertheless, I am back and ready to use this platform for more self-reflection and discovery. Maybe it will also help me see where exactly I need to be doing some more review, and it will help you to find out what I’ve been doing for the past seven months. Yay!

So, what have I been doing with all my time, if not trying to come up with phonetic analyses of ventriloquism or explaining the historical development of homophony to my father? The answer, in short, is that I’ve been writing and refining my research proposal. If you tuned in but actually don’t care to read the full essay, that’s all you really need to know. Otherwise, read on to find out more about what that means, and what it is I’m proposing.

For those of you outside academia, the research proposal is exactly what it sounds like: a 20-30 page paper in which I outline what my dissertation will be about and why it’s a topic worth researching. “Here’s what we know,” “here’s what we don’t know,” and “here’s how I plan to find out,” in other words. In my particular program, there is no deadline to complete the proposal, but we’re strongly advised to get it done within the first year, because it helps you establish a trajectory and also helps your supervisor/other professors to see that you’re on the right track. Seeing as I officially started in November last year, I’m more or less right on time.

For the most part, I’ve just been whittling away at my various ideas and cobbling together a coherent roadmap for the remainder of my studies. Back in June, I gave a presentation at my department’s research colloquium to get feedback on the initial ideas for my proposal, and then in July I started drafting. Now I’m finished with the first draft and waiting for feedback from my supervisor, and then I’ll rework it and submit it to the university.

What I’m Actually Proposing

As you may recall, my dissertation topic deals with recursion, which in linguistics, is the idea of embedding a syntactic structure inside itself. For example, in an utterance like Wendy’s boyfriend’s ship’s captain, each instance of [noun]’s is embedded in another; thus, you have to figure out what/who Wendy’s boyfriend refers to before you can figure out what Wendy’s boyfriend’s ship refers to, and so on. The types of structures that allow recursion vary across languages, and are relatively challenging for children/foreign language learners to acquire.

People have looked at the acquisition of recursion in a number of languages including English, Mandarin, Tamil, Japanese, Spanish, and German, and have proposed a number of different explanations for why it’s so difficult for children and language learners. One interesting aspect of this prior work is that the age of acquisition – that is, the age at which most children can reliably interpret recursion like an adult – varies pretty extensively, from as young as four to as old as seven. This variation is not only seen across languages, but also within languages. I would hazard a guess that this has to do with the experimental methods and differing definitions of “acquisition,” but there could also be some important differences between the structures and languages themselves. These potential differences are one thing that I am interested in, and that I want to look at for my doctoral work.

The other major question that my research deals with is the potential impact of acquiring two languages at once on acquiring recursion, particularly when those two languages differ structurally. This taps into a major focus of contemporary bilingual research, which is the degree to which a bilingual child’s two languages interact with one another throughout the course of development. In the past 2-3 decades, we have found that bilingual children’s development does tend to differ from monolingual children both qualitatively and quantitatively, but the extent of and the contexts in which these differences occur have yet to be determined.

There have been three prior studies on bilingual children’s acquisition of recursion compared to monolinguals, each with a different result: one study (Leandro and Amaral, 2014) showed that the bilingual children did better on a comprehension task compared with monolinguals of the same age. Another study (Avram et al., 2021) found that seven-year-old bilinguals performed less accurately than seven-year-old monolinguals in a different comprehension test. And finally, Pérez-Leroux et al. (2017) found no difference in the frequency with which bilingual and monolingual children used recursive phrases to describe pictures. In other words, being bilingual could accelerate, delay, or have no effect at all on the acquisition of recursion. 🤷‍♀️

Of course, a few things that may have impacted the outcomes of previous work, which I hope to control for in my own experiments. Perhaps most important is that each of the aforementioned studies looked at a different language pair: the first one looked at English and Wapichana, the second at Romanian and Hungarian, and the third at English and Spanish. Therefore, characteristics specific to those languages may have influenced the way that bilingual kids interpret or produce recursion. One possibility that was mentioned in the L+A and the Avram studies is whether the target languages differed in terms of branching direction. In the L+A study, where the bilinguals seemed to outperform monolinguals, both languages had left-branching possessives. On the other hand, in the Avram study, where the bilinguals did not do as well as the monolinguals, the children’s two languages had conflicting branching directions for the target structure. This led the authors of both studies to propose that structural similarities could either accelerate or delay bilingual children’s acquisition of recursion.

This proposal regarding recursion taps into a more general hypothesis by Hulk and Müller (2000), who suggest two criteria for determining whether a particular structure will be vulnerable to cross-linguistic influence in bilingual development:

1. The structure is not “purely” syntactic, but involves a connection between syntax and some other module of grammar;
2. There is a (perceived) degree of overlap between the two languages

If a particular phenomenon in language meets both of these requirements, they argue, then it is likely that the bilingual child will show evidence of one language influencing the other. Setting aside the first criterion, #2 predicts that, if a bilingual child encounters a recursive structure such as Wendy’s boyfriend’s ship’s captain in both their languages, they might use evidence from one language to support their development in the other.

Alternatively, the child might speak a language where the branching directions differ: imagine that in Language A, the “head” of Wendy’s boyfriend’s ship’s captain is “captain,” like in English. But in Language B, the head of the phrase is actually “Wendy,” and everything else is a modifier! This is confusing, and may cause the child to incorrectly analyze the structure in one or both languages. They might decide to go with the “head = last” approach, as in Language A, and therefore misinterpret the meaning of the phrase in Language B, or vice versa. They could also decide to just try both interpretations in both languages, resulting in a lower success rate in both. Of course, ultimately, they will sort it all out, but the path they take to get there is what interests me.

I’m working on the design for an experiment now that will address the comprehension of two types of recursion by children acquiring English and Mandarin Chinese at the same time. The two types of recursion differ in terms of branching direction: possessive structures, such as Wendy’s boyfriend’s ship’s captain, are left-branching in both English and Mandarin, with the head of the phrase at the end. Locative structures, on the other hand, are left-branching in Mandarin, but right-branching in English. For example, an expression like the cinema next to the post office next to the bank in English is a mirror image in Mandarin: 银行旁边的邮局旁边的电影院 bank next-to post office next-to cinema. This might constitute confusing evidence for a child, who might prefer to assume that all structures in both languages are left-branching, and therefore may have more trouble acquiring recursive locatives in English.

Thus, if there is an impact of branching direction on the acquisition of recursion specifically, then we would expect the children to perform similarly on possessives (which have the same branching direction in both English and Mandarin) but not on locatives (which are right-branching in English but left-branching in Mandarin). This is what I hope to figure out using a comprehension “game.”

That’s the bare-bones sketch of my current plan and related activities. Last week, I’ve spent most of my time reading up on theories of bilingualism and cross-linguistic influence, and will continue doing so this week. After I have re-worked my proposal and gotten the approval from my supervisor, then I will start ironing out the details of the experiment, which is a fun but time-consuming process that probably warrants its own blog post(s). And then, at some point after that, I will be ready to start recruiting participants.

The good news is that I feel like I know what I’m doing more each day. Stay tuned to see what happens next!