TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Language is not essential for the cognitive processes that underlie thought

562 点作者 orcul7 个月前

55 条评论

dang7 个月前
All: please don&#x27;t comment based on your first response to an inevitably shallow title. That leads to generic discussion, which we&#x27;re trying to avoid on HN. <i>Specific</i> discussion of what&#x27;s new or different in an article is a much better basis for interesting conversation.<p>Since we all have language and opinions about it, the risk of genericness is high with a title like this. It&#x27;s like this with threads about other universal topics too, such as food or health.
Animats7 个月前
This is an important result.<p>The actual paper [1] says that functional MRI (which is measuring which parts of the brain are active by sensing blood flow) indicates that different brain hardware is used for non-language and language functions. This has been suspected for years, but now there&#x27;s an experimental result.<p>What this tells us for AI is that we need something else besides LLMs. It&#x27;s not clear what that something else is. But, as the paper mentions, the low-end mammals and the corvids lack language but have some substantial problem-solving capability. That&#x27;s seen down at squirrel and crow size, where the brains are tiny. So if someone figures out to do this, it will probably take less hardware than an LLM.<p>This is the next big piece we need for AI. No idea how to do this, but it&#x27;s the right question to work on.<p>[1] <a href="https:&#x2F;&#x2F;www.nature.com&#x2F;articles&#x2F;s41586-024-07522-w.epdf?sharing_token=1BwycCwx1wQ2Sfnub1-o0NRgN0jAjWel9jnR3ZoTv0MYYopL5qENCL5gCQ3HDKyBWf6AQLs-HC3fMMzU9skb40K1DK-HWblYUyHTAQuuliWeLXeg5lXVNFOTa3fVek1R0et9kPjIgQljFd2wX1hSlqWjpOKSrRjz8t2mUDQ6Vr6DlhIlAndISxjxnRU2FPd2XMQFK5UDTh5Osiq6IYOksvy1nGE68d0y9YuJvr4Zrok%3D&amp;tracking_referrer=www.scientificamerican.com" rel="nofollow">https:&#x2F;&#x2F;www.nature.com&#x2F;articles&#x2F;s41586-024-07522-w.epdf?shar...</a>
评论 #41892068 未加载
评论 #41894031 未加载
评论 #41891228 未加载
评论 #41893400 未加载
评论 #41893534 未加载
评论 #41890104 未加载
评论 #41891639 未加载
评论 #41891063 未加载
评论 #41895908 未加载
评论 #41890470 未加载
评论 #41891262 未加载
评论 #41892518 未加载
评论 #41905326 未加载
评论 #41892576 未加载
评论 #41895796 未加载
评论 #41897835 未加载
评论 #41891749 未加载
评论 #41893732 未加载
评论 #41892642 未加载
评论 #41893748 未加载
评论 #41892738 未加载
评论 #41891507 未加载
评论 #41896479 未加载
评论 #41897059 未加载
评论 #41896476 未加载
评论 #41897270 未加载
评论 #41894713 未加载
评论 #41893960 未加载
评论 #41896452 未加载
评论 #41896512 未加载
评论 #41892137 未加载
评论 #41897757 未加载
评论 #41893555 未加载
评论 #41892603 未加载
评论 #41891383 未加载
fjfaase7 个月前
As some who has a dis-harmonic intelligence profile, this has been obvious for a very long time. In the family of my mother there are several individuals struggling with language while excelling in the field of exact sciences. I very strongly suspect that my non-verbal (performal) IQ is much higher (around 130) than my verbal IQ (around 100). I have struggled my whole life to express my ideas with language. I consider myself an abstract visual thinker. I do not think in pictures, but in abstract structures. During my life, I have met several people, especially among software engineers, who seem to be similar to me. I also feel that people who are strong verbal thinkers have the greatest resistance against idea that language is not essential for higher cognitive processes.
评论 #41892035 未加载
评论 #41891193 未加载
评论 #41892129 未加载
评论 #41907543 未加载
评论 #41891928 未加载
评论 #41893304 未加载
评论 #41896097 未加载
评论 #41894530 未加载
YeGoblynQueenne7 个月前
&gt;&gt; They’re basically the first model organism for researchers studying the neuroscience of language. They are not a biological organism, but until these models came about, we just didn’t have anything other than the human brain that does language.<p>I think this is completely wrong-headed. It&#x27;s like saying that until cars came about we just didn&#x27;t have anything other than animals that could move around under its own power, therefore in order to understand how animals move around we should go and study cars. There is a great gulf of unsubstantiated assumptions between observing the behaviour of a technological artifact, like a car or a statistical language model, and thinking we can learn something useful from it about human or more generally animal faculties.<p>I am really taken aback that this is a serious suggestion: study large language models as in-silico models of human linguistic ability. Just putting it down in writing like that rings alarm bells all over the place.
评论 #41895496 未加载
orwin7 个月前
I will add an anecdata, then ask a question.<p>I could enter what we all here call the &quot;Zone&quot; quite often when i was young (once while doing math :D). I still can, but rarely on purpose, and rarely while coding. I have a lot of experience in this state, and i can clearly say that a marker of entering the zone is that your thoughts are not &quot;limited&quot; by language anymore and the impression of clarity and really fast thinking. This is why i never thought that language was required for thinking.<p>Now the question: would it be possible to scan the brain of people while they enter the zone? I know it isn&#x27;t a state you can reach on command, but isn&#x27;t it worth to try? understand the mechanism of this state? And maybe understand where our thought start?
评论 #41891620 未加载
评论 #41892210 未加载
评论 #41900274 未加载
orobus7 个月前
I’m not a neuroscience expert, but I do have a degree in philosophy. The Russell quote immediately struck me as misleading (especially without a citation). The author could show more integrity by including Russell’s full quote:<p>&gt; Language serves not only to express thoughts, but to make possible thoughts which could not exist without it. It is sometimes maintained that there can be no thought without language, but to this view I cannot assent: I hold that there can be thought, and even true and false belief, without language. But however that may be, it cannot be denied that all fairly elaborate thoughts require words.<p>&gt; Human Knowledge: Its Scope and Limits by Bertrand Russell, Section: Part II: Language, Chapter I: The Uses of Language Quote Page 60, Simon and Schuster, New York.<p>Of course, that would contravene the popular narrative that philosophers are pompous idiots incapable of subtlety.
评论 #41896852 未加载
评论 #41896381 未加载
Geee7 个月前
The important question is: what is considered a language?<p>&gt; You can ask whether people who have these severe language impairments can perform tasks that require thinking. You can ask them to solve some math problems or to perform a social reasoning test, and all of the instructions, of course, have to be nonverbal because they can’t understand linguistic information anymore.<p>Surely these &quot;non-verbal instructions&quot; are some kind of language. Maybe all human action can be considered language.<p>A contrarian example to this research might be feral children, i.e people who have been raised away from humans.[0] In most cases they are mentally impaired; as in not having human-like intelligence. I don&#x27;t think there is a good explanation why this happens to humans. And why it doesn&#x27;t happen to other animals, which develop normally in species-typical way whether they are in the wild or in human captivity. It seems that most human behavior (even high-level intelligence) is learned &#x2F; copied from other humans, and maybe this copied behavior can be considered language.<p>If humans are &quot;copy machines&quot;, there&#x27;s also a risk of completely losing the &quot;what&#x27;s it like to be a human&quot; behavior if children of the future are raised by AI and algorithmic feeds.<p>[0] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Feral_child" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Feral_child</a>
ryandv7 个月前
It&#x27;s worth noting the precise and narrow sense in which the term &quot;language&quot; is used throughout these studies: it is those particular &quot;word sequences&quot; that activate particular regions in the brain&#x27;s left hemisphere, to the exclusion of other forms of symbolic representation such as mathematical notation. Indeed, in two of the studies cited, [0] [1] subjects with language deficits or brain lesions in areas associated with the &quot;language network&quot; are asked to perform on various mathematical tasks involving algebraic expressions [0] or Arabic numerals [1]:<p>&gt; DA was impaired in solving simple addition, subtraction, division or multiplication problems, but could correctly simplify abstract expressions such as (b×a)÷(a×b) or (a+b)+(b+a) and make correct judgements whether abstract algebraic equations like b − a = a − b or (d÷c)+a=(d+a)÷(c+a) were true or false.<p>&gt; Sensitivity to the structural properties of numerical expressions was also evaluated with bracket problems, some requiring the computation of a set of expressions with embedded brackets: for example, 90  [(3  17)  3].<p>Discussions of whether or not these sorts of algebraic or numerical expressions constitute a &quot;language of mathematics&quot; aside (despite them not engaging the same brain regions and structures associated with the word &quot;language&quot;); it may be the case that these sorts of word sequences and symbols processed by structures in the brain&#x27;s left hemisphere are not <i>essential</i> for thought, but can still serve as a useful psychotechnology or &quot;bicycle of the mind&quot; to accelerate and leverage its innate capabilities. In a similar fashion to how this sort of mathematical notation allows for more concise and precise expression of mathematical objects (contrast &quot;the number that is thrice of three and seventeen less of ninety&quot;) and serves to amplify our mathematical capacities, language can perhaps be seen as a force multiplier; I have doubts whether those suffering from aphasia or an agrammatic condition would be able to rise to the heights of cognitive performance.<p>[0] <a href="https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;17306848&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;17306848&#x2F;</a><p>[1] <a href="https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;15713804&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;15713804&#x2F;</a>
mmooss7 个月前
A concept in every human culture - i.e., created in every culture, not passed from one to some others - is <i>mentalese</i> [0]: &quot;A universal non-verbal system of concepts, etc., conceived of as an innate representational system resembling language, which is the medium of thought and underlies the ability to learn and use a language.&quot; [1]<p>If you look up &#x27;mentalese&#x27; you can find a bunch written about it. There&#x27;s an in-depth article by Daniel Gregory and Peter Langland-Hassan, in the incredible Stanford Encyclopedia of Philosophy, on <i>Inner Speech</i> (admittedly, I&#x27;m taking a leap to think they mean precisely the same thing). [2]<p>[0] Steven Pinker, <i>The Blank Slate: The Modern Denial of Human Nature</i> (2002)<p>[1] Oxford English Dictionary<p>[2] <a href="https:&#x2F;&#x2F;plato.stanford.edu&#x2F;entries&#x2F;inner-speech&#x2F;" rel="nofollow">https:&#x2F;&#x2F;plato.stanford.edu&#x2F;entries&#x2F;inner-speech&#x2F;</a>
smallerfish7 个月前
A little late to the thread, but this is obvious if you&#x27;ve done any reasonably serious mindfulness practice. When you are meditating, you can get to the point where the internal monolog (the yabbering of the &quot;crazy monkey mind&quot;) is completely silenced. &quot;You&quot; are still present, and can direct your attention, and can observe all of the perceptions with full comprehension, without the verbal layer interpreting for you.
评论 #41896499 未加载
评论 #41895271 未加载
psychoslave7 个月前
&gt;You can ask whether people who have these severe language impairments can perform tasks that require thinking. You can ask them to solve some math problems or to perform a social reasoning test, and all of the instructions, of course, have to be nonverbal because they can’t understand linguistic information anymore. Scientists have a lot of experience working with populations that don’t have language—studying preverbal infants or studying nonhuman animal species. So it’s definitely possible to convey instructions in a way that’s nonverbal. And the key finding from this line of work is that there are people with severe language impairments who nonetheless seem totally fine on all cognitive tasks that we’ve tested them on so far.<p>They should start with what is their definition of language. To me it&#x27;s any mean you can use to communicate some information to someone else and they generally get a correct inference of what kind of representations and responses are expected is the definition of a language. Whether it&#x27;s uttered words, a series of gestures, subtle pheromones or a slap in your face, that&#x27;s all languages.<p>For the same reason I find extremely odd that the hypothesis that animals don&#x27;t have any form of language is even considered as a serious claim in introduction.<p>Anyone can prove anything and its contrary about language if the term is given whatever meaning is needed for premises to match with the conclusion.
评论 #41890078 未加载
评论 #41890279 未加载
评论 #41895280 未加载
评论 #41890114 未加载
bmacho7 个月前
There is a <i>non-verbal me</i>. E.g. who moves my limbs, feels the feelings (hunger, happiness, ..), and sometimes helps my <i>verbal me</i> to think (in math or in chess the answer just appears for the verbal me), or in sudden situations it takes over, and it makes decisions very fast.<p>Since it controls my limbs, I consider it to be the <i>real me</i>. My inner monologue is a bit frustrated that it can&#x27;t control my limbs, and it can&#x27;t really communicate with whoever controls my limbs.<p>Then there is my inner monologue, which does my thinking almost always, in an auditory way: imagine <i>the sound of</i> spoken words in an ~5 sec long duration, and let the answer appear. I consider it as an auditory deducing thingy, and also an intelligence on its own.<p>I am mostly fine with this, tho I am curious about my non-verbal me, and I wish I&#x27;d know more about it.
评论 #41894654 未加载
m4637 个月前
I like Temple Grandin&#x27;s &quot;Thinking the Way Animals Do&quot;:<p><a href="https:&#x2F;&#x2F;www.grandin.com&#x2F;references&#x2F;thinking.animals.html" rel="nofollow">https:&#x2F;&#x2F;www.grandin.com&#x2F;references&#x2F;thinking.animals.html</a>
Tagbert7 个月前
I’ve been hearing&#x2F;reading about people who don’t have an inner monologue. Their experience of cognition is not verbally-based.<p><a href="https:&#x2F;&#x2F;www.cbc.ca&#x2F;news&#x2F;canada&#x2F;saskatchewan&#x2F;inner-monologue-experience-science-1.5486969" rel="nofollow">https:&#x2F;&#x2F;www.cbc.ca&#x2F;news&#x2F;canada&#x2F;saskatchewan&#x2F;inner-monologue-...</a>
评论 #41893938 未加载
评论 #41892833 未加载
aniijbod7 个月前
Thought and language are intertwined in ways we don’t fully grasp. The fact that certain cognitive tasks, like comprehension, can proceed without engaging traditional language-related brain regions doesn&#x27;t mean thought doesn&#x27;t use language—it just means we might not yet understand how it does. Thought could employ other forms of linguistic-like processes that Fedorenko&#x27;s experiments, or even current brain-imaging techniques, fail to capture.<p>There could be functional redundancies or alternative systems at play that we haven&#x27;t identified, systems that allow thought to access linguistic capabilities even when the specialized language areas are offline or unnecessary. The question of what &quot;language in thought&quot; looks like remains open, particularly in tasks requiring comprehension. This underscores the need for further exploration into how thought operates and what role, if any, latent or alternative linguistic functionalities play when conventional language regions aren&#x27;t active.<p>In short, we may have a good understanding of language in isolation, but not necessarily in its broader role within the cognitive architecture that governs thought, comprehension, and meaning-making.
评论 #41894693 未加载
评论 #41896596 未加载
评论 #41891481 未加载
shsbdncudx7 个月前
When we eventually nail agi, I think we will look at llm’s as nothing more than the interface to ai, how we interact with it, but we won’t consider it to be ai.
Peteragain7 个月前
I know a little about this area and there is certainly a movement (glacial) away from thinking that thinking uses symbols, distributed or not. The argument cannot be made in a popular science article and so such articles inevitably fall back on popular ideas of what thinking is. The alternatives: the embodied nature of reasoning is one direction and many talk of an &quot;enacivist&quot; approach. There are certainly some kinds of thinking that require symbols, but a surprisingly large and diverse range of intelligent behaviour can be done by just wiring stuff up. Interestingly, a significant amount seems amenable to a mechanism based on &quot;glorified auto-complete&quot; (cf Hinton) and I have written something on the sociological variant - something readable I hope - arxiv.org&#x2F;abs&#x2F;2402.08403
andai7 个月前
When I was 13 or so, a friend asked me, &quot;So, you speak three languages. Which one do you think in?&quot; and the question left me speechless, because until that moment I hadn&#x27;t considered that people think in words. It seemed a very inefficient way to go about things!<p>Much later, I did begin to think mostly in words, and (perhaps for unrelated reasons?) my thinking became much less efficient.<p>Also related, I experienced temporarily enhanced cognition while under the influence of entheogens. My thoughts, which normally fade within seconds, became stretched out, so that I could stack up to 7 layers of thought on top of each other and examine them simultaneously.<p>I remember feeling greatly diminished, mentally, once that ability went away.
评论 #41892296 未加载
adrian_b7 个月前
I can think without language about all the things that I have experienced directly through some of my senses, but there is a huge number of things that I have never experienced directly and about which I can think only using language.<p>I doubt that this is different for other people. I believe that those people who claim that they never think using language are never thinking about the abstract or remote things about which I think using language.<p>For instance, I can think about a model of CPU without naming it, if it has been included in some of the many computers that I have used during the years, by recalling an image of the computer, or of its motherboard, or of the CPU package, or recalling some experiences when running programs on that computer, how slow or how responsive that felt, and so on.<p>I cannot think about a CPU that I have never used, e.g. Intel 11900K, without naming it.<p>Similarly, I can think without language about the planet Jupiter, which I have seen directly many times, or even about the planet Neptune, which I have never seen with my eyes, but I have seen in photographs, but I cannot think otherwise than with words about some celestial bodies that I have never seen.<p>The same for verbs, some verbs name actions about which I can think by recalling images or sounds or smells or tactile feelings that correspond with typical results of those actions. Other verbs are too abstract, so I can think about the corresponding action only using the word that names it.<p>For some abstract concepts, one could imagine a sequence of images, sounds etc. that would suggest them, but that would be like a pantomime puzzle and it would be a too slow way of thinking.<p>I can look at a wood plank thrown over a precipice and I can conclude that it may be safe to walk on it without language, but if I were to design a bridge guaranteed to resist to the weight of some trucks passing on it, I could not do that design without thinking with language.<p>Therefore I believe that language is absolutely essential for complex abstract thinking, even if there are alternative ways of thinking that may be sufficient even most of the time for some people.
评论 #41893891 未加载
zmmmmm7 个月前
Knowing someone with a brain injury, something that is hugely apparent is how much we take for granted &quot;sequencing&quot; - that is, the ability for the brain to hold a sequence of events, ideas or actions in a coherent order over a period of time. It&#x27;s much more fragile than you would think. People with specific brain injuries suddenly can&#x27;t work out whether to put their shoes on before their socks etc.<p>Why I mention this is that I see both language and reasoning as rooted in this more fundamental cognitive ability of &quot;coherent sequencing&quot;. This sits behind all kinds of planning and puzzling tasks where you have to project forward a sequence of theoretical actions and abstractly evaluate the outcome.<p>Which is all to say, I don&#x27;t think language and reasoning are <i>the same</i>, but I do think it is likely they stem from the same underlying fundamental mechanisms in our brain. And as a consequence, it&#x27;s actually quite plausible that LLMs can reconstruct mechanisms of reasoning from language, in a regressive model kind of fashion. ie: just because their are other ways to reason doesn&#x27;t exclude language as a path to it.
评论 #41894457 未加载
ilaksh7 个月前
One thing that always seemed important to these discussions is that the serial structure of language is probably not an optimization but just due to the reality that we can only handle uttering or hearing one sound at a time.<p>In my mind there should be some kind of parallel&#x2F;hierarchical model that comes after language layers and then optionally can be converted back to a series of tokens. The middle layers are trained on world models such as from videos, intermediary layers on mapping, and other layers on text, including quite a lot of transcripts etc. to make sure the middle layers fully ground the outer layers.<p>I don&#x27;t really understand transformers and diffusion transformers etc., but I am optimistic that as we increase the compute and memory capacity over the next few years it will allow more video data to be integrated with language data. That will result in fully grounded multimodal models that are even more robust and more general purpose.<p>I keep waiting to hear about some kind of manufacturing&#x2F;design breakthroughs with memristors or some kind of memory-centric computing that gives another 100 X boost in model sizes and&#x2F;or efficiency. Because it does seem that the major functionality gains have been unlocked through scaling hardware which allowed the development of models that took advantage of the new scale. For me large multimodal video datasets with transcripts and more efficient hardware to compress and host them are going to make AI more robust.<p>I do wish I understood transformers better though because it seems like somehow they are more general-purpose. Is there something about them that is not dependant on the serialization or tokenization that can be extracted to make other types of models more general? Maybe they are tokens that have scalars attached which are still fully contextualized but are computed as many parallel groups for each step.
keepamovin7 个月前
When I was in junior high, I remember a friend saying to me “you can’t think in images, you think in words.” She insisted, and couldn’t believe that I actually thought in images a lot of the time. she was pretty smart and creative.<p>But I thought in images and I still do in part. so I don’t think you need words to think.<p>I thought the people who did were overly computerized, maybe thinking in an over defined way.
wnmurphy7 个月前
I think the argument is in whether &quot;thought&quot; only applies to conscious articulation or whether non-linguistic, non-symbolic processes also qualify.<p>We only consciously &quot;know&quot; something when we represent it with symbols. There are also unconscious processes that some would consider &quot;thought&quot;, like driving a car safely without thinking about what you&#x27;re doing, but I wouldn&#x27;t consider those thoughts.<p>I find an interesting parallel to Chain of Thought techniques with LLMs. I personally don&#x27;t (consciously) know what I think until I articulate it.<p>To me this is similar to giving an LLM space to print out intermediary thoughts, like a JSON array of strings. Language is our programming language, in a sense. Without representing something in a word&#x2F;concept, it doesn&#x27;t exist.<p>&quot;Ich vermute, dass wir nur sehen, was wir kennen.&quot; - Nietzsche, where &quot;know&quot; refers to labeling something by projecting a concept&#x2F;word onto it.
talkingtab7 个月前
A <i>very</i> long time ago I took a programming aptitude test, supposedly from IBM. The test was essentially detecting pattern anomalies. Two straight lines, one crooked. Pick the crooked. The patterns became increasingly more complex. I remember a little voice in my head verbalizing &quot;two straight, one crooked&quot;. But at some point the voice stopped but I was sure which item broke the pattern.<p>My take away is that language is secondary to thinking - aka intuitive pattern detection. Language is the Watson to Sherlock.<p>The corollary is that treating language as primary in decision making is (sometimes) not as effective as treating it as secondary. At this point in my life (I&#x27;m old) I seem to have spent much of my life attempting to understand why my pattern matching&#x2F;intuition made a choice that turned out to be so superior to my verbal language process.
kensai7 个月前
Wasn’t this known at least empirically for centuries? I mean obviously persons and animals without language capabilities (uneducated, deaf, mute) manage some cognitive processes that underlie thought. They might not be the brightest, but it’s there.<p>I guess this was the experiment the proved the point.
ajb7 个月前
One interesting corollary of this is the need to rethink the underpinnings of therapy. Eg, CBT is based around verbal thoughts and replacing bad ones with good ones. I&#x27;ve had CBT practitioners insist to me that thoughts <i>always</i> include words. But once you recognise that there are kinds of thinking, both processing and &quot;mental actions&quot; , not linked to words, it&#x27;s not so easy. How do you identify and replace a maladaptive mental process, if it&#x27;s not linked to a verbalisation? If it is, does replacing the verbalisation really do anything?<p>This I think is why so much popular psychology is so vacuous - the slogans are merely things that triggered some people to figure out how to improve their mental actions, but there&#x27;s no strong linkage between the two.
gibsonf17 个月前
The key to human intelligence are concepts. We just use whatever language we choose to symbolize the concepts.
ziofill7 个月前
I&#x27;m wondering about the &quot;non-verbal language&quot; that scientists use to communicate with people affected by aphasia. What makes a brain with aphasia understand it? Do brains have dedicated circuitry to process words? (as opposed to, say, sounds which are a more general concept)
WiSaGaN7 个月前
I think we need to distinguish between the language e.g. the native language the person uses like English and the concept of language. Your information exchanging binary messages over PCI bus is also part of a language.
mannyv7 个月前
Imo just like in computers, language can make certain thoughts easier to think.
necovek7 个月前
While getting confirmation of this relationship (or lack of it) is exciting, none of this is surprising: language is a tool we &quot;developed&quot; further through our cognitive processes, but ultimately other primates use language as well.<p>The one thing I wonder is if it&#x27;s mostly &quot;code duplication&quot;: iow, would we be able to develop language by using a different region of the brain, or do we actually do cognitive processes in the language part too?<p>In other words, is this simply deciding to send language processing to the GPU even if we could do it with the CPU (to illustrate my point)?<p>How would one even devise an experiment to prove or disprove this?<p>To me it seems obvious that our language generation and processing regions really involve cognition as well, as languages are very much rule based (even of they came up in reverse: first language then rules): could we get both regions to light up in brain imaging when we get to tricky words that we aren&#x27;t sure how to spell or adapt to context like declensions of foreign words<p>&gt; But you can build these models that are trained on only particular kinds of linguistic input or are trained on speech inputs as opposed to textual inputs.<p>As someone from this side of the &quot;fence&quot; (mathematics and CS, though currently obly a practicing software engineer), I don&#x27;t think LLMs provide this opportunity that is in any way comparable to human minds.<p>Comparing performance of small kids developing their language skills (I&#x27;ve only had two, but one is enough to prove by contradiction) to LLMs (in particular for Serbian), LLMs like ChatGPT had a much broader vocabulary, but kids were much better at figuring out complex language rules with very limited number of inputs (noticed by them making mistakes on exceptions by following a &quot;rule&quot; at 2 years of age or younger).<p>The amount of training input GenAI needs is multiple orders of magnitude larger compared to young kids.<p>Though it&#x27;s not a fair comparison: kids learn language by listening, immitation, watching, smelling, hearing and in context (you&#x27;ll talk about bread at breakfast).<p>So let&#x27;s be careful in considering LLMs a model of a human language process.
karaterobot7 个月前
&gt; British philosopher and mathematician Bertrand Russell answered the question with a flat yes, asserting that language’s very purpose is “to make possible thoughts which could not exist without it.” But even a cursory glance around the natural world suggests why Russell may be wrong.<p>I don&#x27;t know why Russell is catching strays. Saying language exists to make possible thoughts which could not exist without it does not in any way imply that you can&#x27;t think without language.
stevebrown7 个月前
Language plays a role similar to that of paper and pen in solving certain math problems. As a tool, it aids deeper thinking. It serves two key functions: facilitating communication and enhancing thought processes. This is why &quot;chain of thought&quot; type of intermediate language prompts improve reasoning in OpenAI&#x27;s o1 model.
jjtheblunt7 个月前
The conclusion implied by the title seems self evident for anyone who has seen any (at least) nonhuman mammalian predator.
评论 #41894854 未加载
评论 #41891072 未加载
评论 #41892310 未加载
bassrattle7 个月前
Is this the death of the Sapir-Whorf theory?
评论 #41889785 未加载
评论 #41889743 未加载
评论 #41892823 未加载
airstrike7 个月前
<a href="https:&#x2F;&#x2F;archive.is&#x2F;PsUeX" rel="nofollow">https:&#x2F;&#x2F;archive.is&#x2F;PsUeX</a>
kaiwen17 个月前
Here&#x27;s what Helen Keller had to say about this in _The World I Live In_:<p>&quot;Before my teacher came to me, I did not know that I am. I lived in a world that was a no-world. I cannot hope to describe adequately that unconscious, yet conscious time of nothingness. I did not know that I knew aught, or that I lived or acted or desired. I had neither will nor intellect. I was carried along to objects and acts by a certain blind natural impetus. I had a mind which caused me to feel anger, satisfaction, desire. These two facts led those about me to suppose that I willed and thought. I can remember all this, not because I knew that it was so, but because I have tactual memory. It enables me to remember that I never contracted my forehead in the act of thinking. I never viewed anything beforehand or chose it. I also recall tactually the fact that never in a start of the body or a heart-beat did I feel that I loved or cared for anything. My inner life, then, was a blank without past, present, or future, without hope or anticipation, without wonder or joy or faith.<p>It was not night—it was not day.<p>. . . . .<p>But vacancy absorbing space, And fixedness, without a place; There were no stars—no earth—no time— No check—no change—no good—no crime.<p>My dormant being had no idea of God or immortality, no fear of death.<p>I remember, also through touch, that I had a power of association. I felt tactual jars like the stamp of a foot, the opening of a window or its closing, the slam of a door. After repeatedly smelling rain and feeling the discomfort of wetness, I acted like those about me: I ran to shut the window. But that was not thought in any sense. It was the same kind of association that makes animals take shelter from the rain. From the same instinct of aping others, I folded the clothes that came from the laundry, and put mine away, fed the turkeys, sewed bead-eyes on my doll&#x27;s face, and did many other things of which I have the tactual remembrance. When I wanted anything I liked,—ice-cream, for instance, of which I was very fond,—I had a delicious taste on my tongue (which, by the way, I never have now), and in my hand I felt the turning of the freezer. I made the sign, and my mother knew I wanted ice-cream. I &quot;thought&quot; and desired in my fingers. If I had made a man, I should certainly have put the brain and soul in his finger-tips. From reminiscences like these I conclude that it is the opening of the two faculties, freedom of will, or choice, and rationality, or the power of thinking from one thing to another, which makes it possible to come into being first as a child, afterwards as a man.<p>Since I had no power of thought, I did not compare one mental state with another. So I was not conscious of any change or process going on in my brain when my teacher began to instruct me. I merely felt keen delight in obtaining more easily what I wanted by means of the finger motions she taught me. I thought only of objects, and only objects I wanted. It was the turning of the freezer on a larger scale. When I learned the meaning of &quot;I&quot; and &quot;me&quot; and found that I was something, I began to think. Then consciousness first existed for me. Thus it was not the sense of touch that brought me knowledge. It was the awakening of my soul that first rendered my senses their value, their cognizance of objects, names, qualities, and properties. Thought made me conscious of love, joy, and all the emotions. I was eager to know, then to understand, afterward to reflect on what I knew and understood, and the blind impetus, which had before driven me hither and thither at the dictates of my sensations, vanished forever.<p>I cannot represent more clearly than any one else the gradual and subtle changes from first impressions to abstract ideas. But I know that my physical ideas, that is, ideas derived from material objects, appear to me first an idea similar to those of touch. Instantly they pass into intellectual meanings. Afterward the meaning finds expression in what is called &quot;inner speech.&quot; When I was a child, my inner speech was inner spelling. Although I am even now frequently caught spelling to myself on my fingers, yet I talk to myself, too, with my lips, and it is true that when I first learned to speak, my mind discarded the finger-symbols and began to articulate. However, when I try to recall what some one has said to me, I am conscious of a hand spelling into mine.<p>It has often been asked what were my earliest impressions of the world in which I found myself. But one who thinks at all of his first impressions knows what a riddle this is. Our impressions grow and change unnoticed, so that what we suppose we thought as children may be quite different from what we actually experienced in our childhood. I only know that after my education began the world which came within my reach was all alive. I spelled to my blocks and my dogs. I sympathized with plants when the flowers were picked, because I thought it hurt them, and that they grieved for their lost blossoms. It was two years before I could be made to believe that my dogs did not understand what I said, and I always apologized to them when I ran into or stepped on them.<p>As my experiences broadened and deepened, the indeterminate, poetic feelings of childhood began to fix themselves in definite thoughts. Nature—the world I could touch—was folded and filled with myself. I am inclined to believe those philosophers who declare that we know nothing but our own feelings and ideas. With a little ingenious reasoning one may see in the material world simply a mirror, an image of permanent mental sensations. In either sphere self-knowledge is the condition and the limit of our consciousness. That is why, perhaps, many people know so little about what is beyond their short range of experience. They look within themselves—and find nothing! Therefore they conclude that there is nothing outside themselves, either.<p>However that may be, I came later to look for an image of my emotions and sensations in others. I had to learn the outward signs of inward feelings. The start of fear, the suppressed, controlled tensity of pain, the beat of happy muscles in others, had to be perceived and compared with my own experiences before I could trace them back to the intangible soul of another. Groping, uncertain, I at last found my identity, and after seeing my thoughts and feelings repeated in others, I gradually constructed my world of men and of God. As I read and study, I find that this is what the rest of the race has done. Man looks within himself and in time finds the measure and the meaning of the universe.&quot;
swayvil7 个月前
Does the act of assigning meaning to any thing count as language?<p>What if the things are part of a set, chosen for uniqueness and distinguishability. Meanings determined by tradition?<p>There&#x27;s a lot of territory between the two.
评论 #41896743 未加载
upghost7 个月前
Well this comment is about the article not LLMs so I doubt it will have much in the way of legs, but this work has already been covered extensively and to a fascinating depth by Jaak Panksepp [1].<p>His work explores the neuropsychology of emotions WAIT DON&#x27;T GO they are actually the <i>substrate of consciousness</i>, NOT the other way around.<p>We have 7 primary affective processes (measurable hardware level emotions) and they are not what you think[2]. They are considered primary because they are <i>sublinguistic</i>. For instance, witnessing the color red is a primary experience, you cannot explain in words the color red to <i>someone who has not ever seen it before</i>.<p>His work is a really fascinating read if you ever want to take a break from puters for a minute and learn how people work.<p>PS the reason this sort of research isn&#x27;t more widely known is because the behaviorist school was so incredibly dominant since the 1970s they made it completely taboo to discuss subjective experience in the realm of scientific discourse. In fact the emotions we are usually taught are not based on emotional states but on muscle contractions in the face! Not being allowed to talk about emotions in psychological studies or the inner process of the mind is kinda crazy when you think about it. So only recently with neuroimaging has it suddenly become ok to acknowledge that things happen in the brain independent of externally observable behavior.<p>[1] <a href="https:&#x2F;&#x2F;a.co&#x2F;d&#x2F;6EYULdP" rel="nofollow">https:&#x2F;&#x2F;a.co&#x2F;d&#x2F;6EYULdP</a><p>[2] - seeking - fear - anxiety and grief - rage - lust - play!!! - caring<p>[3] if this sounds familiar at all it&#x27;s because Jordan Peterson cites Jaak Panksep all the time. Well 50% of the time, the other 50% is CG Jung and the final 50% is the book of Exodus for some reason.
评论 #41892951 未加载
codersfocus7 个月前
While not essential for thought, language is a very important tool in shaping and sharing thoughts.<p>Another related tool is religion (for emotions instead of thoughts,) which funnily enough faces the same divergence language does.<p>Right now society that calls itself &quot;secular&quot; simply does not understand the role of religion, and its importance in society.<p>To be clear, I don&#x27;t belong to any religion, I am saying one needs to be invented for people who are currently &quot;secular.&quot;<p>In fact, you have the disorganized aspects of religion already. All one needs to spot these are to look at the aspects that attempt to systematize or control our feelings. Mass media, celebrities for example.<p>Instead of letting capitalistic forces create a pseudoreligion for society, it&#x27;s better if people come together and organize something healthier, intentionally.
评论 #41890673 未加载
评论 #41895889 未加载
habitue7 个月前
Language may not be essential for thought, (most of us have the experience of an idea occurring to us that we struggle to put into words), but language acts as a regularization mechanism on thoughts.<p>Serializing much higher dimensional freeform thoughts into language is a very lossy process, and this kinda ensures that mostly only the core bits get translated. Think of times when someone gets an idea you&#x27;re trying to convey, but you realize they&#x27;re missing some critical context you forgot to share. It takes some activation energy to add that bit of context, so if it seems like they mostly get what you&#x27;re saying, you skip it. Over time, transferring ideas from one person to the next, they tend towards a very compressed form because language is expensive.<p>This process also works on your own thoughts. Thinking out loud performs a similar role, it compresses the hell out of the thought or else it remains inexpressible. Now imagine repeated stages of compressing through language, allowing ideas to form from that compressed form, and then compressing those ideas in turn. It&#x27;s a bit of a recursive process and language is in the middle of it.
评论 #41890579 未加载
评论 #41890458 未加载
评论 #41890360 未加载
hackboyfly7 个月前
Well it’s important to note that this does not mean that our language does not play a role in shaping our thoughts.<p>“You cannot ask a question you that you have no words for”<p>- Judea Pearl
评论 #41889596 未加载
评论 #41889595 未加载
fnordpiglet7 个月前
For those who can’t and don’t think in words this is unsurprising.
评论 #41889769 未加载
评论 #41889753 未加载
评论 #41889604 未加载
评论 #41890199 未加载
评论 #41889537 未加载
评论 #41889526 未加载
fsndz7 个月前
more proof that we need more than LLMs to build LRMs: <a href="https:&#x2F;&#x2F;www.lycee.ai&#x2F;blog&#x2F;drop-o1-preview-try-this-alternative" rel="nofollow">https:&#x2F;&#x2F;www.lycee.ai&#x2F;blog&#x2F;drop-o1-preview-try-this-alternati...</a>
SecuredMarvin7 个月前
Thanks, dang.<p>I think that using a LLM as the referred telepathy device to a wolfram-alpha&#x2F;mathematica like general reasoning module is the way to AGI. The reasoning modules we have today are still much to narrow because of the very broad and deep search trees exploding in complexity. There is the need for a kind of pathfinder which could come from common knowledge already encoded in LLMs, like in o1. An system playing with real factual reasoning but exploring in directions coming from world knowledge.<p>What is still missing is the dialectic between possible and right, a physics engine, the motivation of analysed agents, the effects of emergent behavior and a lot of other -isms. But they may be encoded in the reasoning-explorer. And of course loops, more loops, refinement, working hypotheses and escaping cul-de-sacs.<p>There are people with great language skills and next to no reasoning skills. Some of them have general knowledge. If you ever talked to them, for a at least an hour freely meandering topics you will know. They seem intelligent for a couple of minutes but after a while you realise that they can refer fact, even interpret metaphors, but they will not find an elegant one, to navigate abstraction levels, even to differentiate root cause from effect or motivation and culture from cold logic. Some of them even ace IQ or can program but none did math so far. They hate, fear or despise rational results violation their learned rules. Sorry, chances are if you hate reading this, maybe you are one (or my English is annoyingly bad).<p>I love talking to people outside my bubble. They have an incredible broad diversity in abilities and experiences.
farts_mckensy7 个月前
Stix&#x27;s claim appears to be unfalsifiable. In scientific and philosophical discourse, a proposition must be falsifiable—there must be a conceivable empirical test that could potentially refute it. This criterion is fundamental for meaningful inquiry.<p>Several factors contribute to the unfalsifiability of this claim:<p>Subjectivity of Thought: Thought processes are inherently internal and subjective. There is no direct method to observe or measure another being&#x27;s thoughts without imposing interpretative frameworks influenced by social and material contexts.<p>Defining Language and Thought: Language is not merely a collection of spoken or written symbols; it is a system of signs embedded within social relations and power structures. If we broaden the definition of language to include any form of symbolic representation or communication—such as gestures, images, or neural patterns—then the notion of thought occurring without language becomes conceptually incoherent. Thought is mediated through these symbols, which are products of historical and material developments.<p>Animal Cognition and Symbolic Systems: Observations of animals like chimpanzees engaging in strategic gameplay or crows crafting tools demonstrate complex behaviors. Interpreting these actions as evidence of thought devoid of language overlooks the possibility that animals utilize their own symbolic systems. These behaviors reflect interactions with their environment mediated by innate or socially learned symbols—a rudimentary form of language shaped by their material conditions.<p>Limitations of Empirical Testing: To empirically verify that thought can occur without any form of language would require accessing cognitive processes entirely free from symbolic mediation. Given the current state of scientific methodologies—and considering that all cognitive processes are influenced by material and social factors—this is unattainable.<p>Because of these factors, Stix&#x27;s claim cannot be empirically tested in a way that could potentially falsify it. It resides outside the parameters of verifiable inquiry, highlighting the importance of recognizing the interplay between language, thought, and material conditions.<p>Cognitive processes and language are deeply intertwined. Language arises from collective practice; it both shapes and is shaped by the material conditions of the environment. Thought is mediated through language, carrying the cognitive imprints of the material base. Even in non-human animals, the cognitive abilities we observe may be underpinned by forms of symbolic interaction with their environment—a reflection of their material engagement with the world.<p>Asserting that language is not essential for thought overlooks the fundamental role that social and material conditions play in shaping both language and cognition. It fails to account for how symbolic systems—integral to language—are embedded in and arise from material realities.<p>Certain forms of thought might appear to occur without human language, but this perspective neglects the intrinsic connection between cognition, language, and environmental conditiond. Reasoning itself can be viewed as a form of internalized language—a symbolic system rooted in social and material contexts. Recognizing this interdependence is crucial for a comprehensive understanding of the nature of thought and the pivotal role language plays within it.
评论 #41896152 未加载
评论 #41892877 未加载
Razengan7 个月前
Those without, do you feel “jealous” of people with a “mind’s eye”?<p>Or vice versa?
评论 #41892834 未加载
评论 #41892871 未加载
jostmey7 个月前
Progress with LLMs would seem to support the title. The language abilities of LLMs does not seem to lead to higher thought, so there must be additional processes that are required for higher thought or process that don’t depend on language
评论 #41892305 未加载
tekknik7 个月前
oh how many people have been oppressed and suppressed from a misguided understanding of the alternative…<p>who cares right?
psychoslave7 个月前
&gt;And in fact, most of the things that you probably learned about the world, you learned through language and not through direct experience with the world.<p>Most things we know, we are probably not aware of. And for most of us, direct experience of everything that surrounds us in the world certainly exceeds by several order of magnitude the best bandwidth we can ever dream to achieve through any human language.<p>Ok, there are no actual data to back this, but authors of the article don&#x27;t have anything solid either to back such a bold statement, from what is presented in the article.<p>If most of what we know of the world would mostly be things we were told, it would obviously be mostly a large amount of phatic noises, lies and clueless random assertions that we would have no mean to distinguish from the few stable credible elements inferable by comparing with a far more larger corpus of self experiments with realty.
lazyasciiart7 个月前
Now I need to learn about how they convey these questions without language.
acosmism7 个月前
now I really want to understand the deep thoughts my cat is having
评论 #41889928 未加载
nickelpro7 个月前
As always, barely anyone reads the actual claims in the article and we&#x27;re left with people opining on the title.<p>The claims here are exceptionally limited. You don&#x27;t need spoken language to do well on cognitive tests, but that has never been a subject of debate. Obviously the deaf get on fine without spoken language. People suffering from aphasia, but still capable of communication via other mechanisms, still do well on cognitive tests. Brain scans show you can do sudoku without increasing bloodflow to language regions.<p>This kind of stuff has never really been in debate. You can teach plenty of animals to do fine on all sorts of cognitive tasks. There&#x27;s never been a claim that language holds dominion over all forms of cognition in totality.<p>But if you want to discuss the themes present in Proust, you&#x27;re going to be hard pressed to do so without something resembling language. This is self-evident. You cannot ask questions or give answers for subjects you lack the facilities to describe.<p>tl;dr: Language&#x27;s purpose is thought, not all thoughts require language
评论 #41890390 未加载
评论 #41890861 未加载
评论 #41890007 未加载
评论 #41889875 未加载
评论 #41890316 未加载
评论 #41889973 未加载
评论 #41892886 未加载
yarg7 个月前
Is this not obvious?<p>Language is a very poor substitute for freely flowing electrical information - it is evolved to compensate for the bottlenecks to external communication - bottlenecks that are lacking an internal analogue.<p>It&#x27;s also a highly advanced feature - something as heavily optiimised as evolved life would not allow something as vital as cognition to be hampered by a lack of means for high fidelity external expression.
评论 #41892896 未加载
评论 #41890070 未加载
评论 #41894321 未加载
评论 #41892454 未加载
eth0up7 个月前
Considering that, in 2024, if not a majority, then, still, a vast portion of our consciousness is words. Perhaps not for the illiterate, but for many, much of our knowledge is through the written or spoken word. [Edit: Even a hypothetical person, alone and isolated, never having spoken, would still devise internal language structures, at least for the external realm. ]<p>Base consciousness is surely not dependent on language, but I suspect base consciousness may be extremely different from what one might expect, so much that compared to what we perceive as consciousness, might seem something close to death.
评论 #41890858 未加载