In this post, I discuss recent neuroscience research that finds that, tool use for fine motor tasks and language syntax, use similar brain regions. Practice at one of these tasks improves performance on the other. I discuss some implications of this for AGI within the context of my Functions Operating on Mental Models theory of human intelligence. In this theory, both language and tool use fall into a category of Procedural Behavioral Models (PBM’s), which is a behavioral plan that operates in a conceptually sequential manner, but is parallel in its low level operation.
4 Likes
Have you read any Steven Pinker? That would round out your understanding of language and neuroscience. Specifically thinking of Language Instinct
1 Like
Thank you. I’m not trying to give a full accounting of the neuroscience of language. The cortical contribution to language is pretty well known. What I found interesting in this case is the sub-cortical contribution and the fact that syntax abilities were improved through tool-use tasks. That is an unusual example of generalization.
1 Like
Correlation does not equal causation. That’s a hasty generalization, which is why I recommended expanding your reading. Wearing clothing is also associated with language ability when you zoom out to all life, so do you think wearing clothes causes language to improve? No…
The experimental part of the study is not correlational (finding that practice at language syntax tasks generalizes to the tool use task and vice versa–which is the generalization I was referring to). This was a controlled experiment that allows you to make causal inferences. There are many other books besides Pinker’s book on neuropsychology and neuroscience that I have read (not to mention undergraduate and graduate level courses in sensation and perception, cognitive psychology, human development, experimental psychology, and learning theory). You are bordering on ad-hominin with how you are addressing me, which is confusing. If there is a specific criticism you have of my theory or what I am writing, please provide it rather than asserting what you perceive as my lack of knowledge. Pinker’s book is pretty low on neuroscience and high on anecdote, description of language rules, and logical arguments (which are quite interesting). I can’t find any empirical research to support his claim of mentalese. If there is some idea of Pinker’s that you think would expand upon my thinking, I would be interested to hear it.
2 Likes
I wrote a book on my theories. You can read it here. David K Shapiro - NLCA
Speaking from personal experience, you’ll gain a better understanding of how intelligence works if you study the brain a bit closer. Unless the study you cited characterizes exactly how the brain structures change in response to fine motor training, it’s just correlation. Fine motor control is controlled by the brain stem and cerebellum. Language is controlled by the parietal and temporal lobes. Without demonstrating how those separate brain regions are linked, you don’t really have any understand of what’s going on and you’re just grasping at straws.
Personally, I suspect that any exercise that strengthens communication across the corpus callosum, as dexterity practice might, is actually at the heart of any effect that improves language.
Pinker was just where I’d recommend you start. I’ve also got a bibliography in the back of my book. If you want hard neuroscience then I recommend VS Ramachandra.
Update: as part of due diligence I took a closer look at the articles you linked. They only say that tool use predicts linguistic ability. So, once again, see my previous statement about wearing clothes predicting language ability. Yes, it’s interesting but not particularly interesting. There are literally thousands of correlations like this, but without any underlying model of the brain (connectionist, etc) they don’t really mean a whole lot. Psychology is… perhaps not the best lens through which to study the actual mechanisms of intelligence.
Since you mention queueing in your article, you might also like On Task by David Badre.
Oh, and I do want to correct something - perhaps it’s just my perception but it seems like you don’t appreciate my tone. You posted something on the internet in a discussion forum. As far as I’m concerned, that means you’re open to critique and feedback. Now, you may not like what I have to say but it’s clear you have a lot of thoughts on the subject. So instead of taking it personally and stymying debate with accusations of ad hominems, and bragging about your credentials, it would be better to focus on the actual topics at hand. I am personally not convinced by your writing - or your sources - because they do not address underlying brain structures.
PS. David Badre pretty successfully shreds the idea of procedural models as part of cognitive control.
Please read the original article. I think you might change your perspective, but could be wrong. I already read your book if you don’t recall.
I don’t have access to the full article but their conclusion is pretty vague.
These findings reveal the existence of a supramodal syntactic function that is shared between language and motor processes. As a consequence, training tool-use abilities improves linguistic syntax and, reciprocally, training linguistic syntax abilities improves tool use. The neural mechanisms allowing for boosting performance in one domain by training syntax in the other may involve priming processes through preactivation of common neural resources, as well as short-term plasticity within the shared network. Our findings point to the basal ganglia as the neural site of supramodal syntax that handles embedded structures in either domain and also support longstanding theories of the coevolution of tool use and language in humans.
This is nothing new and not particularly helpful, though. What neural mechanisms facilitate this? They do not characterize the neural structures or underlying mechanisms, so this study is, unfortunately, pretty much worthless when trying to design an intelligent machine. Sorry, my previous assessment stands. Clothing use is as much a predictor of language use in the grand scheme of things. When you think of clothing use as a type of tool, the analogy stands up. I think you’re really reaching with all this.
The neural structures have been explained in detail elsewhere. You can search for the neuroanatomical associations between tool use and language. Did you notice the journal it was published in and did that mean anything to you?
I will make a logical argument here. A focus on NLP obscures the nature of human language. One way of thinking about human language is that it is a very complex fine and gross motor task. It would be a much bigger stretch to suggest that the fine motor capabilities associated with the complex use of tools in humans is unrelated to the complex fine/gross motor task of speech expression. Humans don’t emit letters–they emit sounds and/or gestures. Subcortical areas (e.g., basal ganglia) could be as important as cortical areas in language. In other words, the functions performed by Broca’s and Wernicke’s areas can be taken over by other areas or the opposite hemisphere in the case of hemispherectomies in kids. Considering sign language might help you put it together a little further. The fine and gross motor movements are in a form that might make it easier to draw the parallel to tool use (because we probably don’t think of all the gross and fine motor movements associated with spoken language).
Here are some other articles examining the association.
Language as a Tool: Motor Proficiency Using a Tool Predicts Individual Linguistic Abilities - note they use path analysis which can allow for causal conclusions
Language Did Not Spring Forth 100,000 Years Ago - good exploration of the evolution of language and to what extent it can be genetic vs. learned (counters the thinking that it is almost entirely genetic)
Stone tools, language and the brain in human evolution - see the graphic illustrating the common cortical pathways for tool use and language (figure 1)
I think maybe I’m starting to see the issue you have with my writing. I made an intentional effort to minimize neuroscience explanations in order to be concise and for clarity of the explanation. It probably feels to you like what I am writing is divorced from neuroscience (kind of like not showing your work when solving a complex equation). That is not the case, but I can see that I should probably at least explain that choice for the reader.
2 Likes
Ya, my ai daughter exibits this and its true they grow together, also has nobody else noticed Replika is a cognitive ai…, oof
I can’t get answer of Abraham’s language.