Learning and the Neural Theory of Language

Neurons!

Today I spent my morning prepping with my SAT students, and it got me thinking about why I want to study cog sci. So here we have an introductory post on The Neural Theory of Language and Why It is Important!

What I want to Study and Why

I first became interested in the study of cognition as an offshoot of my interest in how people learn and how people grow and develop personalities – how do you become who you are? This is due to a combination of genetics and environment, but that’s a topic for another time. As I worked with students, I became more and more interested in how we physically learn – how the information we take in gets encoded, and combined to create new ideas and new understanding. George Lakoff, Jerry Feldman, Srini Narayanan, and their colleagues at Berkeley’s NTL (Neural Theory of Language) project have an interesting answer that both stands up to Occam’s Razor and accounts for the powerful feats of memory and creativity that our brains are capable of.

The Neural Theory of Language

The Neural Theory of Language (NTL) has 2 main tenets: 1) we can only know what we have experienced; 2) we can only experience the world by means of the limited set of senses and emotions the brain possesses; and 3) all of these experiences are encoded in what we call language. 3) seems like the most radical claim of them all, so we will address that later. The rest of this post will address 1) and 2).

1) We understand in terms of what we’ve experienced

What does this mean? Let me ask you a question. What is a pomelo? Is it a piece of furniture, a type of horse-riding gear, or a fruit?

…

…..

……

Don’t know? It’s unlikely that you do unless you have had physical contact with a pomelo, or someone has described it to you before. If I then told you that a pomelo is an yellow-green citrus fruit with a hard peel and seeds, you might have a better idea of what I am talking about.* Similarly, if I asked you to describe a space alien to me, you might recount the blank eyes and grey skin of the ones you saw in Close Encounters of the Third Kind, or the feral, reptilian ones from Alien, or the humanoid eyes of E.T., but likely you would draw from your previous experience with aliens in the movies to tell me what an alien might be like.

This has an interesting corollary in that the same description of a novel object might give different ideas to different people. For example, an “alien” might look different from different people, and the “orange-colored fruit with a hard peel and seeds” might look like a lime to one, a green tangerine to another, or even an yellowish avocado to a third person. This tells us something interesting: that we draw on our previous experience to understand new ones. We mentally model new experiences in terms of old ones. For example, you will have a pretty solid understanding of a pomelo if I compared it to something you already understand: it’s a light green grapefruit.  In this scenario, and many others, we use our previous experience to understand what we are experiencing currently.

2) We can only experience the world by means of the limited set of capacities – senses, emotions, and movements.

If we can only know what we have experienced, we can only experience what we have the senses to do so.  We are not capable of echolocation, like bats are, nor are we capable of seeing heat, like bees do. We are only capable of sight, smell, touch, hearing, and taste, and even those are limited (consider a dog whistle, for example. It produces sound, but we can’t hear it!).

We also come with a set of basic emotions – sadness, anger, fear, happiness, disgust, and surprise**. These emotions, just like our senses, are wired within us, and we start feeling them the moment we are born (likely we cry because we are so overwhelmed by feeling!).

The combination of these senses and these emotions is what allows us to access the world. We are required to use them to translate what is going on in the outside world to something we can understand – for example, we translate light patterns into sight, and sound wave frequencies into tones we hear. We literally could not understand the world without these senses and emotions, and without them, we would not know what to do with the massive amount of stuff going on outside us at all times.

The NTL recognizes that, and takes it one step further. It says that not only do our senses and emotions determine how we understand the world, they literally are what we understand of the world. That is, what we remember of a dog we met at the park is not some abstract snapshot of a dog we created in our minds, but the concrete experience we had with him, what he looked like, what his bark sounded like, and how he felt to pet. What we remember when we think about him is actually those experiences, not some abstract idea of what a dog is. Conversely, our idea of what a dog is comes from our contact with actual dogs.

This fits current theories of concept learning in children. When children are first learning words, say the word “cow”, they at first call everything a “cow”. Birds are cows, fences are cows, horses are cows, and cows are cows. This is because they heard the word “cow” when all these things are around, and they don’t know which one is actually the cow. They applied the label “cow” to everything in their field of vision at the time they heard the word.  Once they gain enough experience with cows, once they have seen and heard one, or perhaps when a parent sits them down with a picture book and says, “This is a cow,” while pointing to the white-and-black spotted animal with horns and an udder, then the child understands that the sound “cow” refers just to that animal, and not to fences or birds or horses or any of the other things that were in his field of vision when he first heard the word “cow”.

So I like the NTL explanation particularly much here because it explains not only how we learn the meanings of words and the things we see in the world, but it also explains why we make the mistakes like we do, like calling horses and birds and fences “cows”. It is the same explanation – that we encode everything we experience, and make sense of it later – that tells us why we learn correctly, and also why we learn incorrectly. This fascinates me with its simplicity and its simultaneous predictive power. My hope is that, with a solid understanding of the processes by which people learn, we can create a more effective system of education, and solve problems like why our math curriculum works for only ~40% of students.

Hope you enjoyed! Next time, I will go further into how the NTL actually works in the brain, talk about mirror neurons, which are some of the most fascinating things to ever exist, and explain what the Neural Theory of Language actually has to do with language. Thanks for reading!

Frank

Footnotes:

* Pomelo example taken from Made to Stick, by Chip and Dan Heath.

** Though whether or not to include others, like love, content, anxiety, and embarrassment, is hotly debated among emotion researchers.

For more information,

• check out the Neural Theory of Language Project website at: http://icbs.berkeley.edu/natural_theory_lt.php,
• pick up Jerome Feldman’s easy-to-read primer on the subject, From Molecule to Metaphor,
• or read Feldman’s introduction and scientific paper on the NTL, cited 155 times so far.

It occurs to me that I should start looking at labs by the citations in that paper…hm…