Pattern detection and language errors

One thing I’ve been thinking about as T learns new words is how our brains are wired to detect patterns and how useful that is for learning language.

Human brains are AMAZING at detecting patterns – and, thinking about and watching T learn new words, it’s so clear to me how important pattern detection is. For example, T has primarily learned the word “horse” (he doesn’t say it yet, but will make a clip-clopping noise when he sees one in a book) by looking at pictures of horses in books and playing with a toy horse that we have at home. But, all of these horses are different – some are realistic photos, some are cartoonish and colored unrealistic colors, some are more realistic cartoon drawings, etc. So, T has learned to generalize across all these different instances of horses to learn some pattern like “a horse is an object with four legs, a longish neck, and a mane, and it makes a clip-clopping noise.” Thinking about this is kind of amazing to me!

But, sometimes it’s possible to learn a pattern incorrectly – for example, by learning a pattern that is a bit too broad. An example of this might be if T had learned the pattern “a horse is an object with 4 legs and is sometimes brown” – this pattern might lead him to identify a picture of a cow as a horse (which sometimes happens with one particular cow in a book that we have :)).

T frequently makes pattern errors that both amuse and interest me. One of the more humorous ones involves his identification of pictures of my father (T’s grandfather, whom we call “thatha” – “grandfather” in Tamil). My father wears transition-lenses glasses (so they frequently look like sunglasses, even inside). T recently saw a picture of Ray Charles and insistently pointed at the photo yelling “Thatha! Thatha!” I guess T’s “pattern” for his grandfather is an older man who wears sunglasses!

One of the things about T’s errors that interests me is which words he tends to make more “errors” with. I think that he tends to “correctly” use nouns much more than non-nouns, and thinking about this in terms of pattern detection, I think this makes sense. I think that the “pattern” for most nouns is generally easier to deduce than for non-nouns. For example, “ball” is a fairly concrete clear concept, compared to, for example, “up” and “down.” “Up” and “down” are used in so many different contexts – lifting T up and down, picking something up off the floor, going up and down steps, etc, whereas “ball” is basically a round toy (although T will sometimes call fruit like melons balls!).

And, I think that T tends to have more interesting interpretations (and by this, I mean broad!) for when words like “all done” and “bye bye” should be used. Perhaps this is because he’s still trying to learn the “pattern” for these words!

So many new words

T’s language seems to have grown quite a lot in the last week! Just in the past week, he’s started saying approximations of “eyes,” “nose,” “mouth,” “up,” “down,” “ball,” “bump,” and “go.” Of course, his version of these words is often not quite right – he’ll usually get the first consonant right and get close to right vowel after, and leaves off the final consonant. But, it’s clear to me what he’s trying to say (for example, if he’s stabbing me in the eye shouting “AYYYY!”). I made a page here documenting all of T’s words that I hope to update over time!

What’s interesting to me is that T’s “favorite” babbling consonant is definitely “d” – it’s predominantly what he uses when he’s just babbling to himself in the stroller (vocal play), or trying to say stuff to us that isn’t understandable as a clear word to us. But, despite “d” being T’s clear favorite consonant, he actually has more words that are “b” words.

I made a little graph of how the current set of T’s words are distributed in terms of predominant sound type – for example, whether the primary consonant is bilabial (formed with the lips pressed together – like for “b” or “m”), whether the primary consonant is alveolar (with the tongue pressed at the roof of the mouth – like for “d” or “n”), or whether the word (at least as T says it) is mostly vowels. Here’s the chart:

image (1).png

So, as of right now, a higher proportion of T’s words most prominently feature bilabial consonants, although that’s closely followed by alveolar consonants. I’m not sure if T has just been most interested in words that start with “b” (like “bubbles” or “ball”) or if these might be easier for him to learn, since bilabial consonants are very visually salient (it’s easy to see someone’s lips pressed together compared to where their tongue is inside their mouth), and that’s something that we visually emphasize when we say words we’re trying to teach him.

I’ll be interested to see how this pattern changes over time!

Consonant and Dissonant Sounds

T has gotten really into music in the last month or so. Lately, when he hears music he likes, he’ll start dancing. We went out to eat a few weeks ago, and the restaurant had music playing in the background, and T was dancing up a storm in his high chair – it was adorable! And, I especially love when T dances when I play the piano!

T’s evident love of music made me start thinking about babies and qualities of music that they may innately appreciate. A little reading led me to a few studies (for example, this one by Trainor and Heinmiller) that have shown that even infants that are just a few months old prefer pairs of musical notes played together (a musical interval) that are consonant (pleasant sounding) rather than dissonant (harsh or unpleasant). What’s interesting to me about these studies is that babies seem to recognize and prefer musical intervals widely recognized by adults (both musicians and non-musicians) as being consonant, even without much music-listening experience. Here’s a YouTube video that gives examples of consonant (for example, an octave, a perfect fourth, a perfect fifth, etc.) and dissonant intervals (a minor second, a major second, etc.) – the difference between consonant and dissonant intervals is really striking, even if you don’t know the names of the intervals!

The study I linked to above showed that infants prefer listening to consonant intervals rather than dissonant intervals. And, this study (by Sugimoto et al.) showed than even an infant chimpanzee preferred listening to consonant intervals rather than dissonant intervals. So this seems to suggest that there’s something hard-wired in our brains that makes us prefer consonant musical intervals, even if we haven’t heard much music or had any musical training.

So, I decided to test T to see if he has a preference for consonant musical intervals over dissonant intervals! I played different examples of consonant and dissonant sounds for T, both on an iPad and on the piano to see if he had different reactions (this was not the protocol used in any research study, but it was the best I could do at home :)). When I played consonant and dissonant intervals on the piano, I got no noticeably different reaction from T, although, this may have been because he was preoccupied with trying to lick the fan. I then played consonant and dissonant sounds three different times on the iPad. Two of the times, he started smiling when he heard the consonant sounds and reaching for the iPad, and when he heard the dissonant sounds, he turned away from the iPad and even seemed a little visibly distressed (although this may have been because my experiment was running into snack time; I’m beginning to understand why research studies with babies often have a high attrition rate). The third time got no difference in reaction. So, it seem possible that T has a preference for consonant sounds over dissonant ones, but I can’t really be sure based on this.

One thing that piqued my curiosity is that these studies were done in normal hearing babies (and a normal hearing chimpanzee), so I wondered whether people with hearing loss hear consonant and dissonant intervals different than normally-hearing people. I found this study (Tufts, et al.) which showed that people with hearing loss do hear consonant and dissonant intervals differently, and their explanation of why was so interesting to me!

Before I talk about what Tufts et al. found, here’s a quick explanation of why different intervals are heard as consonant or dissonant. The big difference between different intervals is how far apart the two notes are – for example, a minor third is 3 semitones apart (an example is C and E-flat) and an octave is 12 semitones apart. I think that the accepted theory for why an interval sounds consonant is that the component notes of the interval are far enough apart to be easily resolved by the cochlea – one technical way to say this is that the two notes fall into different auditory filters. Here’s a picture I made to go along with an analogy:

intervals.jpg

Imagine you’re rolling balls down a hill into different buckets (I have no idea why you’d be doing this, but just go along with it!). How far apart two balls are at the top of the hill represents a musical interval, and each bucket at the bottom of the hill represents an auditory filter. If two balls fall into the same bucket, the interval composed of those two “balls” (notes) will sound dissonant, whereas if they fall into different buckets, they’ll sound consonant. People with hearing loss are known to have broader auditory filters – that is, the buckets at the bottom of the hill are a lot bigger, so more balls would fall into them. So, based on this theory, you’d predict that people with hearing loss would find closely spaced intervals dissonant that people with normal hearing would find more consonant.

And, to some extent, this is what Tufts et al. found – here are two cool plots (FIGS. 3 and 4):

tufts.jpg

From Tufts, et al. – Top – FIG. 3 of Tufts, et al – Consonance/Dissonance scores for Normal Hearing adults. Bottom – FIG. 4 of Tufts, et al. – Consonance/Dissonance scores for Hearing Impaired adults.

FIGS. 3 and 4 of Tufts show people’s consonance/dissonance ratings for different musical intervals (indicated by the ratio of the frequencies) – as you can see from FIG. 3 – people with normal hearing say that notes with a ratio of 1 (this is the “unison interval,” or the same note!) is very consonant, and then there’s a steep drop where very closely spaced notes (a frequency ratio between 1.0 and 1.1) sounds VERY dissonant (from the analogy above, where balls are landing in the same bucket), and then as the notes get farther apart, the intervals become more and more consonant (analogy – balls more likely to land in different buckets) as the interval approaches an octave (a frequency ratio of 2.0). While the pattern is similar for people with hearing loss (the bottom figure), the curves look slightly different – if you look at the solid line on the bottom figure, the curve is flatter and has a negative peak later, which shows that people with hearing loss generally rated all intervals as sounding less consonant (because the curve is flatter overall), and that the most dissonant intervals to people with hearing loss were ones that were already “recovering” in consonance for people with normal hearing.

This may explain why people with hearing loss often say that music doesn’t “sound right” to them, even if they wear hearing aids or cochlear implants – the musical intervals that people with normal hearing find pleasant and that are therefore used heavily in music may sound harsh and unpleasant to people with hearing loss.

(By the way, the Tufts, et al. study was done with adults with mild-moderate hearing loss; I have no idea what you would find for babies with hearing loss, which I think is a really interesting question!)

New Language Developments!

The article I wrote about here indicated that once babies first start to say words, their language explodes soon after – and I think we are juuust on the cusp of that with T (1 year old!)!

I wrote here that we were considering T’s first word to be “bubble”; since then, he has definitely said “dada” to refer to his dad (in fact, I didn’t even see my husband behind us, and only he realized he was there when T started grinning and shouting “dada”!), “baba” for his bottle, and “buh” when he saw a school bus. I think he is starting to pair me and “mama” (although he frequently says “mama,” I’m not sure he ties the sound with me). I think he also knows that sound a duck makes; often, if I ask him what a duck says, he’ll say “kwa kwa” (a funny sounding “quack quack!”).

And, his receptive language is really taking off too! Now, when we ask him to point to his ears or our nose or mouth, he will do this correctly. He also understands when we tell him to put things in containers (like his toys in a box or laundry in the basket, although he will almost always pull them right back out again!) or to wait before knocking over a tower of blocks.

It kind of seems like understanding the concept of words (like that different objects, people, or actions have unique sounds that go with them) was a big cognitive leap, and now that he is starting to understand that, he’s ready to learn lots of new words!

Conductive vs. Sensorineural Hearing Loss

I found the distinction between conductive and sensorineural hearing loss confusing, so I wanted to write about it. Please note that I’m not an audiologist, so this explanation is just based on my own understanding!

First, here’s a diagram of the ear:

ear.png

When we hear a sound, the sound goes in through the outer ear (the pinna), travels through the ear canal, and vibrates the ear drum. The vibration of the ear drum causes the ossicles to move (these are three tiny bones in the middle ear), which then causes fluid in the cochlea (the inner ear) to move. The moving fluid in the cochlea causes tiny hair cells in the cochlea to bend, and the location of the hair cells that bend indicate the frequency of the sound (how high or low pitched it is). The information from the hair cells then travels to the brain.

A conductive hearing loss occurs when someone has a hearing loss that stems from a problem in the outer or middle ear. Conversely, a sensorineural hearing loss occurs when someone has a hearing that stems from a problem with the hair cells of the cochlea or with the nerve that travels from the cochlea up to the brain.

The vast majority of diagnosed hearing losses are sensorineural, especially in adults. Sensorineural hearing loss can occur when hair cells are damaged due to medication, infection, and especially from exposure to really loud noise (don’t turn your headphone volume up too high! If people can hear your music when you’re wearing headphones, it’s too loud!). In babies diagnosed with sensorineural hearing loss, the cause might be genetic.

Conductive hearing losses are less common, although they do occur pretty frequently in children. A conductive hearing loss means that sound is having trouble reaching the cochlea – this could be due to a malformation of the ear canal, too much ear wax, or, most commonly in children, fluid in the middle ear and/or an ear infection.

Many conductive hearing losses can be treated or will go away on their own – for example, ear wax can be removed and ear infections can be treated. Conversely, a sensorineural hearing loss can’t be cured – once hair cells are gone, they can’t be grown back! (at least not yet. scientists are walking on this!). Hearing aids and cochlear implants don’t fix sensorineural hearing loss – in the case of hearing aids they amplify sounds to stimulate the remaining hair cells, and in the case of cochlear implants, they bypass the hair cells completely to stimulate the nerve that transmits sound information to the brain.

So how do audiologists determine whether a hearing loss is conductive or sensorineural? I think this is tricky, especially with babies! One thing T’s audiologist does is compare his audiograms measured with air conduction with his audiograms measured with bone conduction. Sounds played by air conduction go through the full ear chain – from the outer ear to the middle ear and then to the inner ear. With bone conduction, the audiologist puts a tiny oscillator on T’s mastoid bone, and the vibrations cause a sound at a particular frequency to be played – but this sound bypasses the outer and middle ears and goes right to the inner ear.

By comparing the air conduction and bone conduction audiograms, the audiologist can get an indication of if there’s something wrong with the outer and middle ears. If both the air conduction and bone conduction audiograms show a hearing loss, and the losses are similar, this indicates a sensorineural hearing loss. On the other hand, if the bone conduction and air conduction results are very different from each other, this may indicate a conductive hearing loss. For example, if the bone conduction audiogram results show normal hearing thresholds, this indicates that the inner ear is normal. If, however, the air conduction audiogram shows abnormal thresholds where the bone conduction results are normal, this indicates that although the inner ear is normal, there is a problem with sound getting to the inner ear – that is, there might be a problem with the outer and middle ears – a conductive hearing loss.

I think this is particularly difficult to get a handle on for babies and young children, because they are too young to be able to tell you if they are hearing stuff differently in one ear compared to the other, the way an adult would be able to if they had ear wax build up or fluid in one ear. And, I think conductive hearing losses in particular can fluctuate a lot, especially for babies and children who go to daycare and school and may frequently have colds or ear infections, so that variability just makes this even trickier to nail down!

First Word!

We’re calling it! T’s first word is “bubble,” pronounced “ba-buh” or “ba-bwah.”

I wrote here about how we weren’t sure if T’s attempts at “bubble” counted as a word. Since then, T has started pointing to the bottle of bubbles and shouting “ba-buh!” to get us to blow bubbles. It’s clear that he’s trying to say “bubble,” and he’s matching the first consonant and first vowel, and he’s trying to get us to do something, so we’re counting it as a word!

 

Bribing Baby With Cheerios

I’m going to take a little break today from talking about hearing and language to talk about something totally unrelated that caught my attention this past week.  The Washington Post had this article about a recently published study about babies’ innate sense of morality and using bribery to override that innate morality that fascinated me. (Tasimi, A. and Wynn, K. “Costly rejection of wrongdoers by infants and children.” Cognition, 151, pg. 76-79, 2016 – full study available here!)

Background

In a nutshell, there have been previous studies that have shown that even 1-year old babies have an innate sense of morality – after watching a puppet show, babies are more drawn to a puppet that helps another puppet than to a puppet that hurts or hinders the other puppet. What this study found was that while this is true, you can override this innate preference for the “good” puppet by bribing the baby with graham crackers.

The Study

The researchers first established that babies are more likely to reach for a plate that has more than one graham cracker compared with just one graham cracker.  They did this by doing a “baseline puppet show” where they had 2 puppets each offer the baby a plate with graham crackers. One of the plates had just one graham cracker, and the other had more than 1 (either 2 or 8) – indeed, the babies robustly seemed to reach for the plate offered by the puppet that had more than 1 graham cracker.

They next conducted a little “morality puppet play” for the babies. There were two versions of the play. In both versions, a lamb puppet tried to open a box. In the first scenario (the “good puppet” scenario) – a helpful rabbit puppet helped the lamb puppet open the box. In the second scenario (the “bad puppet” scenario) – an unhelpful rabbit puppet (wearing a different colored shirt than the helpful rabbit puppet) slammed the box lid shut on the lamb puppet. After showing babies both versions, the researchers had the helpful rabbit puppet and the unhelpful rabbit puppet each offer the baby a plate of graham crackers, with the helpful rabbit puppet offering just one graham cracker and the unhelpful rabbit puppet offering more.

The researchers found that when the unhelpful rabbit puppet offered only 2 graham crackers (that is, more than the helpful rabbit, but not MUCH more), the babies took the helpful rabbit’s crackers – they were rejecting the unhelpful rabbit. However, when the unhelpful rabbit offered 8 graham crackers (MUCH more than the helpful rabbit), the babies took graham crackers from the unhelpful rabbit, suggesting that babies were unable to resist the lure of the bad puppet offering them more crackers.

Here’s the main figure from the study:

morality.jpg

With the gray bars, you can see that babies preferred the plates with more than 1 graham crackers, regardless of whether the plate had 2 or 8 – they just cared that it was more than 1. Looking at the black bar on the left, though, you can see that the babies strongly rejected the plate with 2 graham crackers when it was offered by the bad puppet – but when the bad puppet offered 8 graham crackers, they reached for that plate almost as often as when the 8 graham cracker plate was offered by a puppet without behavior issues.

My Replication

Since T (11.5 months) is pretty much the exact age of the babies in the study now, I had to try and replicate this! (this is why people have children, yes? <– totally kidding). Luckily we had a set of puppets at home (that T LOVES), and I used cheerios instead of graham crackers (for crumb issues).

I set up a little curtain area and started my puppet show. I first tried to replicate the baseline finding, to see if T would prefer cheerios from a puppet offering more than 1 compared to from a puppet offering just 1 cheerio. I was very easily able to replicate this – regardless of which of my puppets was offering cheerios or how much more than 1, T robustly reached for the “more than 1” cup.

Then I had to replicate the morality plays. Rather than having my helpful puppet help open a box and my unhelpful puppet slam the box shut, I had my helpful puppet help carry a block and my unhelpful puppet drop a block on the other puppet’s head. (This was particularly disturbing since our puppets are part of a “happy helpers” set – so in this case, I had a doctor puppet dropping a block on a firefighter puppets head. oops.)

I then had my “good puppet” offer 1 cheerio and my “bad puppet” offer 2 cheerios – to my surprise, T reached for the cup with one cheerio! I ended up repeating this two more times – in the end, of the 3 times I tried this, he went for the good puppet (offering 1 cheerio) twice, and the bad puppet (offering 2 cheerios) once.

Finally, I had my good puppet offer 1 cheerio and my bad puppet offer 10 cheerios – I repeated this 3 times, and of those 3 times, T always reached for the bad puppet’s many cheerios.

Based on this, it seems like we might have replicated the results of the study! T did somewhat seem to reject the bad puppet when he only offered 2 cheerios (2/3 times), but reached for the bad puppet when he had lots of cheerios (3/3 times). On the other hand, I’m not convinced that he thought the “bad puppet” was all that bad, since he DID reach for the bad puppet several times to kiss him, and he clapped and laughed when the bad puppet dropped the block on the other puppet’s head. So it’s hard to say. Regardless, T and I both had fun 🙂