Think words, talk pictures

Do we need natural language to think? What came first - the language or the thinking? Is it possible to think through images only? Fascinating testimonies of autistic people about their special thought processes hint at the possibility of non-linguistic thinking

Rotem Hermon, Odyssey

One of the most fascinating and frustrating topics in the study of consciousness is the relationship between language and thought, or in other words, the role that language plays (if any) within the mechanisms of thinking. The subject is fascinating because of the central place that language occupies in our lives, and because of the almost intuitive connection that we tend to assume exists between language and thought. He is frustrated by the elusive circularity of trying to examine thought and language from within themselves.

This topic accompanies the philosophy from its beginning, with at its center a dispute between two main approaches. One approach claims that thought precedes language, and language is used only for communication and expression of thoughts. A second approach holds that language precedes thought, meaning that there is no thought without language, and thinking can be fully explained as a linguistic act.

During the 20th century, the position that holds that language takes an important and central role in thought processes was strengthened. Today, the philosophical discussion tends to reduce the occupation to intentional thinking, that is, thoughts that can be formulated as statements about the world ("Uri thinks that the ball is round"). Such thinking, it is argued, is carried out as a particular kind of language.

There is a debate surrounding the question of what kind of language it is. Is this the language we speak (the natural language), or is it a special language used only for thinking? Without going into the history and depth of the debate, it can be said that the debate is conducted between two main approaches:

The American philosopher Jerry Fodor developed the "Language of Thought Hypothesis" theory (LOTH - Language of Thought Hypothesis), according to which the system on which the thinking processes are carried out is a special internal language, a kind of "machine language" of the brain, which Fodor calls "mentalization" ".

The assumption is that the language of thought is different from spoken language, and this is also how cognitive processes can be explained in creatures that do not speak language (babies and animals). Also, Fodor claims that we need a primary language "burned" in the brain, which will allow us to learn spoken language. So the language of thought is an innate and unlearned language, preceding any other language.

On the other side of the fence is the philosopher Peter Carruthers, who seeks to defend the position that human thinking is carried out in natural (spoken) language. This position relies heavily on introspection - internal observation of the way of thinking. Such an observation as well as evidence emerging from various cognitive studies show that thoughts are internal representations of sentences in natural language.

Crothers does not claim that all thinking is done in language, but only a part of it - conscious intentional thinking. Unconscious processes can be carried out in other ways. According to him, it is possible that thought processes in creatures without language are carried out in an internal language of thought, as well as the process of learning a first language.

On the basis of this primary thinking ability, higher order thinking is built in mature humans. This is conscious thinking, and it can exist through the use of spoken language.

To summarize the differences between the approaches, Fodor says that in order to understand what the image is supposed to represent, we need an explanation in language (what would an image that represents the claim "Obama is the president of the United States" look like? And what would differentiate such an image from an image that represents the claim "Obama is black" "?). Crothers claims that prior and verbal information is needed to convey the full meaning of the concept that is represented in the image (what would a picture of "democracy" or "yesterday" look like? Does the meaning of the concept of "school" boil down to the appearance of the building?)

One picture is not worth a thousand words

Alongside the familiar experience of "inner speech" there is also another common experience - images that come to mind. Both Fodor and Crothers hold the view that visual representations can only be used as a supplementary means that coexists with thinking in language. Pictures cannot take the place of language for the existence of processes of the intentional thinking type. Their main claim concerns the inability of images to carry semantic meaning without the need for language. Is it so?

Interesting evidence emerges from the reports of autistic people, or as they call themselves, AS (people on the autistic spectrum). This evidence presents forms of non-verbal thinking that call into question positions such as those of Fodor and Crothers, and require us to reexamine our conceptions of the relationship between language and thought. The evidence is not much, but even the little evidence that exists provides us with an extraordinary glimpse into the wondrous possibilities of the human mind.

Dr. Temple watches movies

Dr. Temple Grandin was diagnosed with autism as a child. She has a PhD in animal science, and is responsible for the design of a large part of the facilities used to treat animals in the US. Grandin writes a lot about the visual way in which her thinking works, to which she attributes her unique understanding of animals.

In an attempt to understand what makes the way of thinking of autistic people unique, Grandin talked to hundreds of autistic people and divided the ways of thinking she found into three main categories (she also notes that, in addition to being all non-verbal, almost every way of thinking in autism is unique):

Visual thinking: thinking in pictures. The nature of the images varies from person to person (sometimes still, sometimes XNUMXD and moving).
Relative thinking: these people do not see pictures, but patterns and relationships (between numbers, sounds or relationships in space).
Logical thinking: using a private sign system.

Grandin herself describes her thinking as videos that are projected inside her head. The words for her are a second language - when someone speaks to her, the words are translated into images in her mind. On the films she sees, Grandin can activate productive processes to generate new thoughts.

Grandin details the difficulties of a person with visual thinking in dealing with the spoken language. She says that the easiest words to learn are nouns, because they can be directly represented by pictures. The more difficult the word is to visually represent, the more difficult it is to learn. To understand abstract concepts, Grandin needs to translate them into visual symbols. The concept of "honesty", for example, is translated in her to the image of a person who places his hand on a Bible book in a court of law.

Spatial prepositions (like "above" and "below") had no meaning for Grandin until she learned to attach a picture to them (to this day when she hears the word "below" she sees herself under a desk at school during an aerial bomb defense practice). Linking words are a problem for her. As a child, Grandin omitted words like "is" and "the" because they were meaningless to her, and even today the inflection "to be" is meaningless to her.

When she reads, Grandin translates the text into mental movies, but finds it difficult to translate text that does not have a concrete meaning. For example, she sometimes fails to understand philosophy books or articles that deal with abstract future predictions.

General concepts are also a fallacy. The first step in understanding them, says Grandin, is to learn to create categories and adapt to new information. As a child, Grandin distinguished between dogs and cats by their size. The classification failed when her neighbors got a small dog. She had to find a new visual characteristic, and discovered that all dogs have noses of a certain shape.

Tony Langdon is a computer scientist from Australia who is diagnosed as autistic. Langdon describes a different visual way of thinking than Grandin. His thought process contains "objects" - visual elements that represent ideas, concepts or objects; And "actions" - a component that can describe movement, logical transformations or a process.

A thought for Langdon is an image that consists of a combination of objects and actions. He describes his memory as a structured store of experiences. The information in the database is mostly visual and linked in a network of reciprocal links. The information is saved in a "shrunken" form - new information is not saved in its entirety, but if possible as links to information that already exists in the database. Langdon notes that this mode of memory limits him in remembering precise details of experiences.

Software inside the brain

Ronan Gil is an Israeli computer expert diagnosed with Asperger's Syndrome. In the conversations I had with him, Gil describes a way of thinking that uses a system of logical signs, the closest description to which is a programming language.

Can you describe your thought process?

"I can't really describe my way of thinking in words (mainly since this is a kind of translation in itself), but I think that the closest thing that exists in the outside world (the world that is not my interior or outside what is commonly called - mainly out of lack of understanding and ignorance - my bubble) that can To describe my thought process, is logical signs. However, this differs not only between autistics and non-autistics, but also from autistic to autistic.

How would you describe your use of natural language? Does language have a place in thinking or is it used exclusively for external communication?

"Language, in the conventional sense of the word, has no place in my thought process. Every time I need to transfer my thoughts from my world to the outside world (mainly for the needs of communication with others), the products of these thoughts must undergo further processing (similar in essence to translation) into language. The process itself cannot be translated at all. But the products, sometimes yes. It does not matter at all if it is in Hebrew, English, Arabic or any other language.

In the process of translating from an external language to my internal lips, in order to carry out a thought process, there is a translation for words. There is also, although to a very limited extent, translation of complete sentences. But this translation is only done when I am aware that the translation of each word separately can lead to several results that contradict each other. Such a translation is usually limited to sentences which I know, based on experience, do not come to express the intention behind the words themselves. A very prominent example of this is the question "How are you?" In this question there are social elements, which I cannot distinguish, so I am required to translate the entire sentence and not just the words that make it up."

Would you describe your thinking as taking place in a kind of internal language? Does the thinking process include the ability to use basic elements (kind of words) to assemble more complex elements (which are like sentences)?

I don't usually refer to the factors in my thought process as language, except when I am required to talk about them. The closest word to describe the factors in my thought process is programming language. I believe that this is one of the main reasons for the attraction of many autistic people to the world of computers. The closest thing in the daily life of modern man in the outside world to my thought process, is highly complex software.

It is software that has no dedicated purpose other than the product of thought (and the purpose of that specific thought). In my "inner language" there are simple signs that make up more complex signs. Thus it is similar to letters, syllables, words, sentences, paragraphs and so on. It's just that in language I couldn't go much further than that, while in my thought process, it's much, much more complex and long.

The product of thought is very similar to the built-in function RETURN in software languages. It can hold only the final token of the process and it can contain an entire process itself. However, in any case where I need the final mark and I do not remember the process that led to it, I must repeat the entire process in order to use the product of the process. Sometimes, during such repetition, an error is discovered in the initial process that led to a wrong product. Then, when repeating the process, the final product also changes."

Another and somewhat anecdotal evidence can be found in a letter that Albert Einstein wrote to the mathematician Jacques Hemmard in 1945, in which he describes his thinking as visual-symbolic: "Language or words, whether written or spoken, have no role in my thinking process. The mental entities that serve as components in thinking are signs or images, more or less clear, that can be reproduced and combined at will. ... These games of combination are the main characteristic of a fruitful thought, even before any connection to a logical construction in words or any signs that can be communicated to others".

In all these testimonies, the spoken language functions as a second language used for external communication only, and not as a "natural" language. The thinking described, which is certainly conscious and intentional, is conducted in symbolic systems that are distinctly different from spoken language, including visual systems.

The cognitive difficulties accompanying autism (difficulty in creating general concepts and understanding abstract concepts), as well as the advantages (Grandin's ability to plan) are explained by the special nature of autistic thinking. Such an explanation points to a connection between the medium in which thinking is carried out and the cognitive abilities it confers.

The picture that emerges from these testimonies requires us to rethink the relationship between language and thought. A preliminary hypothesis can be sketched, according to which thinking processes do require a representative system on which they exist, but there are different representational systems that can function as a medium for thinking. The most common system is the linguistic system. However, in some cases - such as autism - whether due to a deficiency in the linguistic system or due to a different "wiring" of the brain, another system can function as the representative system of thinking. This could probably be one of the systems available in the brain, such as the visual, auditory, spatial and similar systems. The differences in the representation abilities of the various systems lead to differences in the cognitive abilities that are made possible by them.

It can be assumed that this phenomenon is not unique only to cases such as autism. A large part of thinking is done in language, but most people, to one degree or another, also experience thoughts in images, sounds, spatial relationships, and more. It is very possible that we all take part in the cognitive processes a variety of systems that enable different types of thinking, each with its own capabilities. A better understanding of the thinking possibilities available to us is important not only for intellectual needs, but especially in practical areas such as education. Then we may be able to develop a flexible and pluralistic learning space, which will not focus only on verbal learning, but will give room for a wider range of human thought abilities.

Rotem Hermon is a software engineer with an interest in technology and its effects on man and society. He graduated with a master's degree in the philosophy, information and digital culture program at Tel Aviv University. The full article was published in the June issue of the "Odysseus" magazine.