Thinking and Language

When adults think, we "hear" our thoughts in our head. It's therefore natural to assume that we think in a particular language, and that what we are hearing represents our raw cognitive processes. I believe that both of these assumptions are wrong.

For the purposes of this discussion, I define language as being a high-level symbolic representation of an idea or concept such as English, French, sign language, hieroglyphics, braille, etc.

First of all, babies begin to understand the world, model it, and make predictions about it way before they ever learn a language. They clearly do not need language in order to function in the world. However, language is useful because it allows sharing of ideas and concepts between creatures. If a baby was abandoned alone in a forest, I have no doubt that it would learn to live well in the forest with little if any need for language.

There seems to be a part of the brain that gradually learns how to convert raw cognitive processes into a form that is suitable for communicating with others. This area is active most of the time, and its output is commonly referred to as a "thought". However, it's important to realize that what you hear in your head is really a translation of a raw cognitive process that has already happened. I tend to think of the raw cognitive process (which is language-independent) as the 'true' thought and what you hear as being the 'translated' thought.

Some people only know how to do this for one language, and others know several languages. When a bilingual persons says that they "think" in German, what's really happening is that they "translate their thoughts" into German. The raw cognitive processes are independent of any particular language.

There's plenty of modern research to back up this theory. For example, when they do MRI scans on people who are asked to move their hand, they find that the neurons responsible for moving the hand start to fire about 1/2 second before the person becomes conscious of making the decision. They are conscious of the 'translated' thought "move my hand", but the 'true' thought happened already.

Realizing that the thoughts you hear in your head are really time-delayed translations of true thoughts is weird. But it's also super-cool!

Comments

[Kevin]

Sounds like you are talking about Platonic ideals? The book Anathem has a discussion on this.

[Graham Glass]

Hi Kevin,

I'm not familiar with that concept, but I have the book and will read it soon!

Cheers,
Graham

[Francisco Gutierrez]

Interesting, I had read about the delay between the motor circuits in the brain and the awareness part of it, making us aware that we moved the hand after we actually made the decision to move it, but that could mean that we have automated that action to the point that the executive center of the brain no longer needs to be conscious of that. It's like learning to drive, when we do it for the first time we have to be very conscious of what we are doing, at the conceptual level, full attention, but once we know it consciousness delegates most of it to subconscious processes. For philosophical reasons, I think that we do need language in order to think at the conceptual human level, and the evidence is that babies that grow up among non-conceptual animals never learn to think like humans, they stay at the perceptual level. But once we have created a concept using language, and we have induced a rule relating to that concept (i.e. dictatorship is against me, so I will be angry when I see a dictator, or when a doctor asks me to move my hand I will move it because I know doctors are not a threat and I can trust them) then the rule can fire subconsciously, without conscious mediation, and consciousness only realizes later that the rule fired.

[Graham Glass]

Hi Francisco,

I think we create abstract symbols that are associated with various concepts in the world, but that these symbols are formed independently from to their association with words in what we traditionally call a "language".

In other words, we form an internal symbol associated with "apple" before anyone tells us that the English word for that symbol is "Apple". Ditto for higher-level constructs that represent temporal relationships or reasoning.

So we can represent the concept that "a doctor is not a threat" without knowing anything about a particular spoken language and its symbols for "doctor" or "threat", or its specific syntax for representing a rule.

It's very important to be precise about what we call a language when discussing these things. I think that the brain creates its own native, internal way of representing knowledge and beliefs about the world which are then associated with spoken/written/signed language later on. Whether we should call this internal representation a "language" is tricky, so for now I tend not to call it a language in order to make my point.

I'm not sure whether the native representation that I allude to satisfies your philosophical requirement for a language. If not, I'm interested to hear why!

Cheers,
Graham

[Mark Baker]

Sounds like you're touching on a favourite topic of mine, the Sapir-Whorf hypothesis;

http://en.wikipedia.org/wiki/Sapir–Whorf_hypothesis

I learned about it while investigating whether we should register our son in "French immersion" at a young age, where all subjects are taught exclusively in French. After reading into it we decided to wait a few years until he's pretty much mastered the basics of English.

[Graham Glass]

Hi Mark,

I definitely think that our surroundings affect the way that we represent the world internally, and not just due to the spoken word but also attitudes, emotional reactions, etc.

So although I think that our brains form representations of concepts and ideas spontaneously in the absence of an external language, the presence of an external language would tend to influence/accelerate the formation of certain symbols.

Cheers,
Graham

[Francisco Gutierrez]

The reason I think we need language is that highly abstract concepts are hard to represent without it. We can represent a tree or a dog with just a picture of a memory of an observation of a tree or a dog, but what about justice, beauty, rights, energy, tensors, complex numbers? It would be very hard to induct these concepts without a precise tool to do it with, and the tool we have is language. For some concepts the tool is the language of mathematics, but it is still a conscious language, with precise syntax and rules, which in the end allows us to represent a concept as a single entity that can be manipulated as if it was a concrete.

So if you agree with me that abstract concepts require precision in order to understand, learn, induct with, deduct with, and hold in one's mind, then there are two possibilities. The first possibility is that the precision is provided by our conscious language, with the syntax and grammar that allows us to express the most complex ideas easily. The second possibility is that there is an equally precise and exact mirror of language internally that does all the thinking. But the second possibility does not satisfy Occam's razor, why postulate a second system when there exists already one that can adequately explain the existence of concepts? That's one of my philosophical objections. Also it seems pretty wasteful that evolution would fail to use the communication system for thinking, or the thinking system for communication, whichever evolved first, and instead have two very similar systems coexisting.

But my main philosophical objection comes from the fact that if the subconscious system actually existed, the only difference between the two systems would be the presence of consciousness. So a subconscious system doing all the thinking would mean that free will is nothing but an epiphenomenon. But free will, like the existence of reality and the existence of consciousness, is at the basis of our knowledge, it is an axiom in the sense that it is presupposed and assumed by everything we ever do or say, even in an argument to refute it you are implicitly assuming it. So anything that denies it would have to be rejected on philosophical grounds.

Have you read Ayn Rand's Introduction to Objectivist Epistemology? She presents her theory of concept formation in that book, and she argues that the conceptual faculty depends on language, and in mathematics actually, it's a very interesting book if you haven't read it I highly recommend it. As an aside, have you read the book Snowcrash? Neal Stephenson plays with the idea of an internal language which is independent of our current modern language, one of my favorite books!

[Graham Glass]

Hi Francisco,

You raise some good issues. I will attempt to address them.

First of all, young children are able to represent concepts like "if I scream my mother will come to help me" and to act upon them. They can do this before they know any spoken language. So using Occams' razor, there doesn't seem to be any need to add an additional kind of thinking mechanism that uses spoken/written language when a native one already exists. As the native one develops more sophistication, more mappings are created that allow the native thoughts to be translated accurately, but the sophistication in language is driven by the evolving native ability, not the other away around.

Another nice thing about this approach is that it explains how bilingual people can sometimes "think" in one language, and sometimes in another. Rather than their brain having two different systems for thinking (say, German and French), it just has a single native system that can be translated into any form. The computer science part of me particularly likes this architecture.

There's a distinction between self-awareness and consciousness. I'm not sure which one you're really referring to. When we look at a sunset, we're conscious of it even if we're not hearing any thoughts about it in our head. If we hear thoughts about something in our head, it's the result of our native cognitive processing being focused on a particular issue.

I believe that consciousness is primarily the existence of the native thought processes themselves, not the translated version that you hear in your head. So the native processing isn't really "subconscious" in the normal sense of the word, it's just that you can't "hear" it.

One corollary of this is that a so-called "philosophical zombie" is an impossible construct - but that's another story!

To me, free will just means that your brain makes choices. The native processing system that I postulate would be both conscious and make choices, so there's no problem.

I have both books you mention - I need to start reading them!

Thanks for these conversations, they're really great fun!

Cheers,
Graham

[Francisco Gutierrez]

Yeah, I enjoy this kind of conversation too, it's definitely a much better way to take breaks from coding than playing in Facebook, haha!

So you are making a less sweeping claim than I thought you were, you are saying that the internal language is still driven by free will, I have no philosophical issue with that claim then. The only problem would be my own experience, I am bilingual in Spanish and English, and I can think fine in both languages, but definitely have certain things that I can think more easily in one language than the other, and certain things I can only think in one or the other but not both. For instance, to add, subtract, divide, and multiply, I have to think in Spanish, but to do calculus, linear algebra, and statistics, I have to think in English. When I have to do these things in the other language, I think internally in my comfortable language first, and then translate. Also, the songs I've written in Spanish are about women, and the one song I've written in English is about politics! This would seem to be evidence against your claim.

But actually my own experience has some evidence for your claim as well. When I am thinking about systems, like physics, math, living things, companies, etc. not the formalism and formulas describing the system, but trying to understand the actual systems themselves, I tend to think without language. I actually visualize the system working, and when I am dealing with something more abstract like math, I end up reducing it to an example that can be represented as a 3D physical system and I end up visualizing that. But that only works with certain systems, I cannot think about philosophy without language for instance.

Also, I am looking around trying to think about things and trying to suppress my internal monologue at the same time, and I can observe some lightning fast automatic connections, like intuitions, but not fully formed thoughts. For example seeing the guitar triggers a picture of myself playing it and a desire to do it which instantly becomes a decision to go do it later, all of it without hearing myself think, but I cannot make the intuitions precise enough without using language.

I agree that it's an elegant design from the computer science perspective, but we were designed by evolution in an environment where there was only one language, so I don't see why evolution would have taken that particular path. Evolution is more like a perl hacker throwing things together to make something work than it is like a software engineer building an elegant and extensible system in ruby. It would be cool if you could come up with some sort of test. Something that would be one way if your hypothesis is true, and would be different if it is false, and then test it.

OK, back to rails...

[Graham Glass]

Hi Francisco,

Maybe there needs to be a special web site called "Philosophy for Rails programmers"!

I think your experience with thinking certain things in particular languages makes sense and has a straightforward explanation. The translation system between your internal "true" thoughts to a language would not work the same for each language. For example, if you learnt basic Math from a Spanish text book but complex Math from an English textbook, the mappings will be weighted accordingly. Did you happen to learn them at different times and with different languages? Also, some languages might have more romance in them and therefore be preferred for a romantic song versus a song about politics!

I'm not sure whether it's possible to suppress the translation system without also suppressing the true thoughts. My guess is that you're actually disrupting the lower-levels of your mind, which would result in your inability to perform complex thinking. We really need a way to suppress the translation system without affecting the rest of the mind. There might already be technology that allows that to be done!

One point to note is that a very young child can recognize, say, a slippery object by touch and distinguish it from a smooth object. They can perform reasoning such as "the slippery object is hard to hold". Clearly, they don't have a word for "slippery" and yet they can represent and manipulate the concept. My theory allows that without any difficulty. I don't know how a theory that requires language for thought would deal with that situation.

Cheers,
Graham

[Steve Grand]

Hi Graham,

I do all my best thinking during the hypnogogic state just before I wake up, and I find I have to deliberately "translate" my ideas into language because otherwise they just vaporise as I become fully awake.

Many times in my life I've "caught a fleeting glimpse of a great idea out of the corner of my eye" (to paraphrase Pink Floyd) but "turned to look and it was gone". If I'm successful at putting it into words I can memorise and recall it, and perform further logical computations on it, but this is invariably in an impoverished form compared to the "medium" in which I first thought it.

What that medium is, I find hard to say. Pictures, movements, feelings - it's a whole sensorium of stuff, but it's definitely not language until I translate it.

So I agree with you.

OTOH my first wife's entire mental world is seemingly made up of words - she claims not to see pictures in her head, and words mean more to her than anything. I doubt she actually *thinks* in words though, it's just that language is her dominant medium for becoming conscious of those thoughts. I suspect we all differ in the reliance we place on different modalities.

Mostly I seem to think in dynamics - I can "visualise" feedback loops and characteristic patterns of change, and I can perform dynamical operations on things in my head - a kind of computation. But I know people who would claim their mental world consists mostly of static images. Perhaps we start to favour one modality as children and that choice gets positively reinforced while the others atrophy?

Since you're interested in education another story on that theme comes to mind. A little boy I knew until recently was doing badly at school, and when I tried to test him on his homework (which was on economics - he's only nine!) he just stared blankly at me. But the school is horribly traditional and everything is taught by rote learning. He was expected to learn precisely worded definitions of things without ever really understanding them. So I sat down with him and explained what was going on in an economy in a dynamical sense (which you'll appreciate), and then drew silly little pictures to illustrate each of the terms. He learned it that way with absolutely no problem at all and you could see the light turn on in his head. His teacher had clearly not accounted for the fact that different people have different kinds of minds, and hence need to learn in different ways.

- Steve

[Graham Glass]

Hi Steve,

You start with an interesting observation regarding the translation and memorization process. One explanation is that you needed to concentrate to create a translated version which was then necessary for memorization. An alternative explanation is that you needed to concentrate in order to better organize your "true" thoughts, which were then translated into the words that you then "heard".

My instinct is to prefer the latter explanation, because it would be quite inefficient to have to convert translate ideas into words/images in order to memorize them properly. In other words (pardon the pun), your ability to articulate the ideas were the result of better organizing the ideas themselves. What do you think?

I agree with your explanation of your first wife's experience.

You then bring up another fascinating point; that our "true" thoughts can represent things that have no straightforward translation into words or images. It sounds like you are more aware of this than most people, but it does make sense; words are many times just an approximation of a richer underlying structure/process.

Your last tale is a good one. It once again shows the loose connection between translated thoughts and 'true' thoughts. The child could remember the translated version but had no 'true' version that actually represented the ideas. Words without meaning. When you were able to connect the words to the actual meaning, the child formed the deeper underlying structures for understanding and manipulating the principles.

Cheers,
Graham