Language Learning and Prediction in Infants: UBC Quinn Memorial by Richard Aslin

Every year there is at least one lecture I eagerly await and look forward to, namely the Quinn Memorial Lecture Series. It has been my fifth attendance over the last six years, with my first one being Michael Gazzaniga’s exquisite lecture entitled Free Will and the Science of the Brain. Unfortunately, I had to miss last year’s lecture by Dr Robert T. Knight because the date and time interfered with my work schedule, but I had been luckier this time around.

Although it certainly did not look like it at first. The scheduled talk Learning and Attention in Infants: The Importance of Prediction in Development by Distinguished Scientist Richard Aslin was suddenly canceled and postponed to a later indefinite date. No reason nor future date were given at the time. Could it be that my brain would have to go two years in a row without the much needed and much desired annual adrenaline shot of knowledge?

Thank goodness, my worries were unfounded, and there I was seated upfront with my smartphone in hand to take notes in preparation for this blog post. I did not know what to expect and on paper, at least for me, the topic at hand about how infants learn language and make predictions about their surroundings did not have the similar emotional impact on me as did previous topics and titles of this wonderful series.

There was certainly nothing wrong with the subject, but for me personally, it came about a decade too late as my son has already taken his first steps into the preteen period. More importantly, I did not wish to hear in retrospect how I may have potentially failed him as a parent in terms of language learning and / or behavior.

Yet again for a second time, my worries were entirely unfounded. The talk itself was much more interesting and engaging than I had imagined. First, Dr. Aslin told us how infants learn through auditory statistical learning. This means that they break streams of words, which are for them initially nothing but random sounds, into auditory chunks, hence creating word boundaries.

These boundaries are often signaled in fluent speech with pauses, such as taking breath. These word chunks then are basically processed and analyzed by the infant brain to make predictions. In order to be able to better understand and predict their surroundings, babies have the tendency to listen and pay more attention to novel and infrequent words and tones.

Why? Because by knowing and establishing certain patterns, they can better understand the rules. This is not a case of merely memorizing words but also looking past them for meaning (vocabulary as a symbol of a designated thing / event in the world) and learning about the inherent rules (grammar, sentence structure and appropriate word choices). 

In a way then, it is memorization plus generalization, respectively known as model-free and model-based learning; the latter of which is generally designated as smart, abstract and flexible ways of comprehension, while the former is rather unfairly treated as the opposite of all those epithets.

Certainly, there is also incidental learning. This means that we absorb knowledge and information without particularly focusing on the given stimuli. This type of learning would occur when one is performing a task by also taking in background noise or information in an almost automatic or subconscious manner. 

Evidently, paying attention is much better suited for learning, but even when we do not notice stimuli implicitly, we are still aware of and capable of remembering strands of information around us without having to specifically focus on them.

The manner they tested all of this was very interesting. The researchers inundated infants with random sounds and stimuli, both in terms of nonsense words, i.e. random syllable sounds as well as tones. Babies tended to be interested in new stimuli, but whenever they managed to discern a clear and repeated pattern, they would lose interest.

This occurred because infants were able to predict the next sound, so the sequence did not provide any novel information for them. Once a pattern was established, the baby moved on to something else … unless there was an unexpected result. That is, if they were expecting a given sound to follow, but either it did not, or the pattern was changed, then the element of surprise would warrant and elicit their attention again.

This happens mainly due to the structure of our brain. To clarify this tendency of the brain to create, establish and predict patterns, they conducted an interesting experiment with pairs of tones. The researchers would play the sound of two honking horns beep beep. After repeated exposure, the baby expected them to come in pairs, yet when the researchers omitted the second beep, the brain nonetheless supplied it.

This was discovered by hooking up wires on babies (no worries, this is harmless and painless standard procedure), thereby noting the infant’s brain activity. In other words, when the brain registered the first honk, the second one was immediately supplied by the brain, regardless if it was or was not there!

Since our brains are wired to make sense of our environment in terms of words or tones, we would use top-down processing once a pattern has been established, meaning that the higher structures of the brain would override the lower ones. In the previous experiment, the higher brain regions literally expected the double tone.

Once inferences were made, babies would then allocate attention to new information. This was observed by their behavior and reactions, such as looking longer at unexpected stimuli or looking away from expected, hence “boring” and unstimulating stimuli. 

In that sense, the brain works tandem with behavior, we are able to see connections and patterns and then start looking for them, hence it is the brain structure that is grounded and established first before the behavior sets in and manifests itself.

About 9 months of age, infants start searching for hidden objects because their brain - and with it their imagination - has developed to a state where the infants are capable of doing and perceiving such a thing; by around 18 months, they can produce two-word sentences. This seems to be universal and is caused by brain development growth and changes.

Yet some of the startling, if not downright shocking, finding was in relation to premature babies. It turns out that they can be at a significant disadvantage when it comes to their brain development although this may not be immediately visible or discernible in their behavior.

Put differently, premature babies may act and behave the same way as other infants, but their brain is less developed than their full-term counterparts. That sent shock-waves down my spine as my son was born prematurely! However, if you have premature children, or if by chance, you are one yourself, keep in mind that this is not always the case.

First off, each case or person is different, and it is not necessarily a disadvantage to begin with. In fact, the premature baby may make up for brain development at a later stage. Due to the plasticity of this magnificent and complex organ of ours, the brain can compensate for parts that have not fully developed, even more so at a younger and developing age. We should also note that the studies were conducted with babies that were significantly premature by about a handful of months and not by a mere month as it was in the case of my son.

Moreover, there are two other general factors that are significant and essential for learning as well. One of them is the fact that salience and prior knowledge tend to drive attention. Anything that is surprising and simply, or maybe ostentatiously, stands out will draw the attention of infants, an observation that is probably equally true for the adult age. The commonplace, however, is generally not worth a second look.

The other factor of importance is what is generally known and referred to as the Goldilocks effect. This simply means that if the information or stimuli presented is too simple or too complex for the baby, he or she will simply look away and lose interest. 

The Goldilocks effect plays also a significant role when it comes to stress and anxiety of children at school; optimal attention and learning is usually achieved when the material and / or environment is neither too comfortable and relaxing nor too stressful and exacting. The middle ground, i.e. the Goldilocks effect, is usually optimal for learning. 

But another question that arose was why was it that we as adults lose that infant ability to make sense of our environment? For instance, this type of processing information would be most useful when learning a second or additional language. Why was the same process not supplied to us at a later stage since it would make our language learning – and life - so much easier?

Part of the problem stems from the fact that as adults we have already established preconceived and set ways of learning and of reacting to our environment, commonly known as entrenched learning. Since we can make more and better sense of our surrounding, and we already have a plethora of prior information and knowledge to select from, we can predict it much better. As a result, and for better or worse, we are not so much drawn to new stimuli and information, but, in a sense, we lose some of our capacity for curiosity and wonder. 

But this is perhaps not the only reason we become somewhat jaded as adults. The other driving force, an issue that came up during my personal conversation with Dr. Aslin afterwards, was anxiety. When we are young, we are generally driven by our anxiety to make sense of everything that is around us, as it could spell potential threat and danger to our health and wellbeing. Once we have sorted out the information, we somewhat lose or at least soften that anxious edge.

With less anxiety, there is also less need to fear or worry about new stimuli. Dr. Aslin called this the dichotomy of an exploring baby versus an exploiting adult brain. While as infants, we strive to look for clues and knowledge to make predictions, as somewhat “jaded” adults, we want to use whatever new knowledge we get our hands on to better serve our benefit and purposes. Most of these developments may origin in the brain and are hence automatic and not necessarily within our control.

This has repercussions in terms of language learning as well. Our life does not so much depend upon making sense of the world since we have already more or less successfully passed through that stage in our younger years. However to finish on a more positive note, we can (and I would say should), despite our brain and age, preserve a sense of wonder by occasionally feeding the child within us and hence ensuring that this worldview or way of interpreting the world is still kept alive and well.