Hacker News Reader

Just as a counterpoint: I (think I) absolutely do think in a specific language. I don't know if I do so all of the time -- there could be times when I don't and don't realize it -- but most of my conscious thought does definitely take the form of words, and sometimes even a well-formed internal dialog.

I speak a second language fluently, and after spending several weeks immersed in that second language I'll sometimes notice that my internal speech switches to it as well, especially when I'm thinking about concepts that I've verbalized/discussed in that language.

Perhaps you are actually thinking in concepts, and in order to help organise and 'sequentialise' those thoughts, you translate them to words in your head at the same time. Therefore, you 'hear' the words, but don't notice the concepts.

We need language really to exchange thoughts in one head to another head, in a slightly richer manner than an ape can.

To "think" we don't really need language (the oral kind the non- deaf think they are thinking in).

Sign language among the born deaf, if allowed to develop without a need to learn spoken languages of the non-deaf, would be clear evidence that thought underlies oral language. And words (that stuff produced for your ear) need not be the only way to think.

> I (think I) absolutely do think in a specific language. I don't know if I do so all of the time -- there could be times when I don't and don't realize it -- but most of my conscious thought does definitely take the form of words, and sometimes even a well-formed internal dialog.

You are mistaken. Your internal monologue is a side effect of your thoughts; it is not your thoughts themselves.

I feel the same way. I always think in language and can't conceive of thinking productively another way. Sure, I can imagine a shape, but that's not something that I'd consider really thinking.

Additionally, for me to really understand some concept, I often have to concretely put it into language so I end up talking to myself. I got a cat so I feel like less of a crazy person - I talk to him and not myself.

Can you think mathematically? Do you know how to code? I would call that thinking and at least for me, that's not verbal. I can explain what I'm thinking but by default I just think without a need to have a voice in my head.

This. Internal dialogue can be mistook to be the only form of thought, but thought itself can be abstract beyond words. Of course the distinction, and the discussion here, might all just be a matter of semantics.

Good point, when I have been programming for a while, and someone asks me a question, it can take a bit for me to get back into human-language mode.

while it has pretty much all my life been for me the same as you describe, i still remember that before and during the first 1 or 2 years of elementary i wasn't thinking in language. And reading - i was reading whole paragraphs as an image instead of sequential word-by-word "reading aloud inside my head" as i've been doing since then. I guess it is about left hemisphere taking control. It is like classical CPU in the left and very powerful GPU in the right.

> We don't think in terms of language. Language is simply a translation layer for the underlying neurological phenomenon. It feels like we think in language because we tend to understand concepts after they've been turned into words.

It always surprises me how many people assume the opposite. Since I don't speak or use my native language during the day, sometimes people will come and ask me whether I "think" in my native language, or in my adopted language. And they're always baffled when I say "neither", because that's not how I think people think.

(Sometimes I, too, may construct internal dialogues in either language, but I never form thoughts themselves as sentences. In fact, more often than not I forget the word that expresses a particular idea, or feeling, or something. That doesn't mean I won't still be able to think about it)

I guess one can think in language although it's not the main way we do it but it is the most observable and you can write it down easily unlike the non verbal but more important parts.

Yes.

"thinking" for me is a stream of thoughts.

But speaking helps to linearize those thoughts and makes them more distinct.

I often just talk to people about my thoughts because it helps meto structure them with the help of language. Sometimes I even get to the point where I have a thought that doesn't map to a word.

I think that thinking can happen in different modes depending on what is useful for the domain you're contemplating. When I am programming I often "see" programs as geometric or organic interlocking structures. I find that I don't really understand a program until I can visualize its structure. On the other hand sometimes I use language and dialog to think out a problem. I remember a specific incident where I was laying in bed on a hot day and I thought "I'm laying here constantly sweating" and then I thought "sweating sounds like a Chinese name" and that's when I realized I had been thinking in Chinese where sweating is pronounced "chu han". You could argue that language-style thought is "translated" from deeper concepts, but I think that goes without saying because there must be some layer of thought that is pre-representation.

Language is a huge part of people's cultural perspective, and it causes all sorts of stupid things to happen. People will say things like "animals don't have morality", when in fact they only fail to have a human morality that humans can communicate. They'll say "animal X doesn't have emotions" - based on what? The fact that they can't communicate with you and demonstrate a human-like response?

Then there's the whole board of culture antagonism built upon the premise of people who can't speak your language being 'lesser people' in some manner.

    > In the future, we'll become more strongly aware of the 
    > limitations semantics places on articulation. People tend 
    > to get too hung up on what words mean, and discussions     
    > will often digress to the moors of semantic posturing.

I'm curious why you think this will be so in the future? Mo matter how much we learn about cognition, the basic process of how people think hasn't changed as far back as we've paid attention. Awareness of the process of thought isn't something that most people ever give attention to beyond idle internal philosophical debates.

Related to this: Anyone else ever focus on their language-based thoughts while thinking them and notice how every mentally verbalized thought seems to have almost a pre-echo of that thought that seems like words but isn't?

Alternatively, my brain is playing a trick on me and I'm filling this sense of echo in after the fact of the language-thought. I've never been able to be discern the sequence with absolute certainty, so figured it was worth putting out there to see if I'm alone in it.

> Anyone else ever focus on their language-based thoughts while thinking them and notice how every mentally verbalized thought seems to have almost a pre-echo of that thought that seems like words but isn't?

Yes. I noticed this when I was first getting into programming and concluded that the verbal part wasn't strictly necessary and should be avoided when possible. Seems to be an issue similar to verbalizing while reading or not (i.e. 'subvocalization'. I always did verbalize initially, but now I can fairly often get into a mode where it feels like the book is reading itself and I understand the words, but there is no enunciation).

I have two main, perhaps conflicting hypotheses on language and thinking now:

1) It's used to solidify trains of thought, similarly to writing down ideas—makes them more stable and easier to return to. Sometimes it's useful to echo your thoughts verbally in order to get this solidification, but it's not necessary.

2) Everything that appears in consciousness is a representation of something deeper going on in the brain. Certain subconscious processes are represented as what we call language in the brain, but the language itself isn't 'active' in any sense.

edit: fixed phrasing.

Semantics is a difficult problem. Wittgenstein once stated "Semantic atoms exist, and I will find them". 20 years later he had to admit defeat and said "Tell me, what's the essence of a chair?". Not long after that, he abandoned rationality completely.

> Anyone else ever focus on their language-based thoughts while thinking them and notice how every mentally verbalized thought seems to have almost a pre-echo of that thought that seems like words but isn't?

Feels like goal-directed behavior, where the function to maximize is "how best can I internally broadcast this exceptionally salient combination of memories and emotions?"

Language basically seems to have two main evolutionary purposes: The first one is the obvious one that we use it to share a common concept space with one another. The second one I am not so sure about, but it seems that language also greatly enhances our thinking. We can exploit the phonological short-term memory in shape of the inner monologue essentially as a scratchpad to efficiently keep track of recent thoughts, and grammars are a great way of combining concepts, so that we can just plug different concepts into the same grammatical structure and obtain new thoughts that might still be grounded in reality due to intuition about the way concepts are typically combined. Moreover, we seem to be able to better verify the soundness of our thinking using language. Oftentimes a thought becomes much clearer when encoded in language. Perhaps grammatical structures are very efficiently linked to causal experiences such that we can very swiftly debug language-coded meanings. The efficient code might also reduce the cognitive load such that we process longer chains of reasoning at a time.

I think language is essentially hacking our memory model which developed around emotional associations with the world. We have sensory input which we translate into associations like danger or pleasure and need to store those in a way that relates in useful ways to the complexities of the way the world works so that we can function as animals. Something like a shriek which might signal danger in one situation might signal happiness in another.

For most people, words start out as sounds, which we are primed by biology to remember and react to. We then form relations around these sounds and their associated symbols or textures for the blind (which sensory input is primary doesn't really matter) and reinforce our associations with emotions and other words to form complex relationships. When we have formed enough of these conections, we can start to establish complex logical arguments around their arrangement and connections. Logical problems become clear in the same way that we are able to react to dangers or opportunities in the real world through the application of memory associations.

So when we string a series of words together, we aren't just bringing up the direct meaning of the sentence. We're bringing into working memory all the various levels of association which those words bring. We're always thinking in these multiple dimensions, whether or not we're directly aware of it. Words are just ways to direct the flow of these associations based on our memories and experiences. When I'm thinking about a sentence, I tend to try to focus more on the periphary thoughts and emotions which they bring to mind rather than focusing on the base meaning.

Our memory is not a model, it is the actual memory we refer to when we talk about memory models. :) Moreover I don't think emotions play a critical role here in the sense that I could imagine a neural architecture without any emotions that still has language.

The idea you've mentioned last is basically what I mean by the scratchpad: We can kind of start associating and explore peripheral thoughts based on a sentence and fairly reliably return the the exact thought from a while ago since our phonological memory is so good.

What I mean by memory model is the way our brain handles access to information and structures the flow of thought. Thought can exist without emotion, and likely does in many living things, but I think the way we structure language is specifically related to our capacity for emotion and the way we attach abstract emotional states to sensory stimulation. I would consider emotion the precursor to symbolic thought in the specific way we do it.

>I believe that language is in the process of deeply integrating into the brain and still isn't really done yet. People have tricks that they use in order to increase the compression rate of words. Narrative for instance. But we still have trouble reaching for words when we need them. I think the language part moves in parallel to the rest of the brain, silently translating thoughts and presenting them to the executive function. (what you think of as you) Thoughts are multi-dimensional, whereas words have many fewer.

Yeah I don't know about all this though. Maybe the brain is very messy and we can't really understand something consciously until we can write/speak it in words, and that words connect disparate thoughts into something meaningful. Maybe this is what you were getting at with 'multi dimensional', but I would think language creates structure, not reduce structure.

For things directly related to the body and rote repetition, the brain might do things we can't consciously express while still being good at it, but that's not really knowledge, that's more automatic neural response. To have knowledge you should have to encode it into some language I think.

On a more broad level, I wonder what symbols like the alphabet and words really 'are', I would tend to think they have no more priority than other stimuli from the environment, but that rather in themselves as an environment for the brain, they have a lot of innate capability for representation and that the brain has learned to manipulate that environment with great skill, but that as far as internal thought goes it's just another environment like navigating a forest or sleeping in a hidden place under the trees to hide from predators. "it's all aesthetic as stimuli" so to speak

> To have knowledge you should have to encode it into some language I think.

What we seem to do is to build concept 'maps' of relating information. The conscious mind has ready access to these maps and uses them to experience the world. I think because most people just aren't terribly intellectual, individual elements in these maps and the very structure of the maps reflect other people around them, what we call culture.

But when we think, we remix and match the map against itself, building up connections between concepts. The more you do this, the less it resembles other peoples'.

A person's definition of a word can change over time. The same basic concept can be represented in crazy-different ways in different people. What I think of when I hear 'Microsoft', is bound to be way different than what someone else thinks.

Be careful to avoid the typical mind policy. Others may use lexicography in their thinking process in a different manner than you are familiar with.

Or your own self-reflection may not be the best guide to what really is happening with your thoughts. Or there might not even be a specific thing happening beyond one's self reflection.

I often have trouble expressing my thoughts concisely for complex concepts. I know the concepts and have a model in my mind of the 'thing' I want to explain. When thinking through, i'll abstract out many things, simply referring to them as 'this' or 'that' in my mind as i try form the sequential narrative to explain them. If a spend too much time trying to think what 'this' and 'that' are, i forget where i am up to in the sequential narrative, needing to start again. I find writing it down helps, but i'll sometimes forget what i'm trying to write half way through it.

> Thoughts are multi-dimensional, whereas words have many fewer.

Another way i like to explain that, is that thoughts are concurrent, while words are sequential. I can think of multiple things at the same time, but i can only speak one word at a time.

There's a quote which i'm not sure where its from, but it something to the effect of:

> Words are power. When you can put a word to a concept, you have power over it.

The correct single word can instantly convey multiple concepts more quickly than many other multiple words can. This enables much better compression and easier conversion of multiple concepts faster and more easily, assuming the recipient of the word is familiar with it. Using multiple words requires the recipient to sequentially construct concepts, leading to the greater possibility of confusion.

I previously never cared much about using the correct words all the time, as it required a lot of effort to remember them (at least on my part). However, i'm beginning to think that the advantages of doing so would be worth the effort.

It's like[1] serializing program state to a sequential form (eg xml, json, protocol buffer, etc).

The sequential form is useful for communication and persistent storage.

While the sequential form can be used for processing, it's clunky, because not its main usage. Consider XSLT.

The set of valid messages is even formally called a language.

[1] Or, rather, serialization is like it.

To really understand this point, I highly recommend playing the board game Concept. It is both really simple and fun, and a deeply thought-provoking intellectual excercise.

https://m.youtube.com/watch?v=k9JWV_HcsqQ

Your first two sentences seem to imply that thought is a neurological phenomenon, but in the following paragraphs you describe several examples of thinking without ever mentioning neurons or any other (one would think) necessary components of neurological phenomena. That's pretty interesting.

Language is a lossy codec for thought.

Lossy, but powerful.

Which is why it has been so successful at co-evolving with and (nearly) subjugating our brains to the extent that it has.

I was speaking to precision, not utility.

> We don't think in terms of language.

What does a thought consist of if not language?

> Language is simply a translation layer for the underlying neurological phenomenon.

It's that... plus a long time formalizing useful terms and concepts, among many other things that fall under "language".

There was somewhere is Daniel Dennett's "Consciousness explained" something in the line of:

- Do you think in words or images?

- I think in thoughts!

These are thought experiments, not coherent arguments.

I would define a hierarchy of reflex < habit < intuition < 'thinking'. Though they all reflect some neurological activity, the only explicitly verbal level is 'thinking', and the others can be conducted unconsciously.

There may be more usefully discriminable levels of abstraction than those. Where to draw the line between thought and unthought? It seems an uninteresting question to me. Should I draw a line between the United States of America and the people thereof? They are distinct, but the latter is a part of the former, and without their aggregate the former would not exist.

Usually I equate thought and consciousness, but perhaps there is a useful distinction there?

In mammals sentience precedes consciousness precedes sapience. In machines the necessity of the linkage is not clear. We can imagine sapience without consciousness, at least, but consciousness without sentience seems more questionable.

We don't think in terms of language.

How do you know?

Beautifully written.

You keep saying "we", you mean "you". Not everyone thinks that way.

You seem very smart https://www.reddit.com/r/iamverysmart