As I write this from my bed in the stroke ward of a very advanced hospital, I think your speculations might have gone too far, not because, as one might imagine, I think the math and science saved me, but because they didn’t.
Contemplating neuroplasticity, my work as a high school English teacher and the potential for my recovering my feet, my balance, and my life, I am sensing a sublime calculus in network math that neither Postman nor Arendt nor Steiner knew or considered.
I’ll have more to say about this later, but the gap you and they sense between speech and ineffable mathematics might be summarized as the gap between the simulated model and the lived experience.
And one of those models is the 20th century, distinctive mindset. Arendt’s distinction, for example, between public and private establishes a foundation for her ongoing argument in The Human Condition. But read the last 30 or 40 pages of that book with a literary ear, and you can hear her feeling loss. It’s downright poetic. Thus, too, Steiner.
Our new world is a networking world, one where mathematical models that operate in neural networks help to prepare us for the strokes to come: climate blasts of major seaports, solar flares, that knock out electrical grids, all kinds of losses of infrastructures and institutions that we created on the foundation of critical distinctions like “public and private.”
Those distinctions can be modeled so that we can reinvent them, and the institutions on which we built our way of being together.
Mathematical modeling, especially with networks and network math, helps us to prepare to heal. There is much more to say on this, and I understand and apologize for the abstraction of that position. But my literal physical position, flat on my back and served by the neural networks that allow me to dictate this post response into my phone, convince me that our artistic heritage (think, say, of Russian formalist Viktor Šklovsky, arguing that by making language strange, poetry helps us live again), including the flow or apparent flow of free speech, has much still to offer us as we render that flow in mathematical models.
Thank you, Abby, for your good wishes as well as this remarkable conversation link. It answers to my intuitions, philosophical and otherwise, and does a fine job of suggesting a framework for the kind of thesis my comment was clawing to articulate. Do you write a blog? I’d love to follow it.
Paul, Thank you for this. I'm sorry to hear of the stroke, but do hope your recovery is progressing. There's much too think about in what you've written and given your response to Abby below, I'd like to take the exchange between Gilman and Bratton into consideration as well. I confess I've had that exchange and several essays that have appeared on Noema bookmarked for some time. I hope to maintain a posture of critical openness to more hopeful outcomes, to be prepared to be surprised as Illich might have put it.
Not arguments, just a few thoughts that occurred to me while reading this:
1) My experience of my own brain is that there is a lot happening in there that is also non-verbal and unseen, so I'm not quite sold on the idea that words deserve quite so much primacy. And between humans, too, there's many a slip. We were never communicating perfectly with words, even when that was all we had.
2) As a writer of fiction, I feel keenly aware of the limitations of words. There are ways I feel and experiences I've had that words are absolutely inadequate for. I would even say that the most important things (to me) are the hardest to convey in words. Language is a blunt instrument. I love it, and will keep whacking away with it, but there's a reason we need music and paintings and dance as well.
3) One of my favourite things is when I'm on one side of the kitchen table, writing or drawing, and my husband and son (a physicist and engineer, respectively) are on the other side of the table working calculus problems together. This is as cozy a father-son bonding moment as I have ever witnessed, as incomprehensible as their mouth-noises are to me. So I don't quite buy that the language of math is only ever cold and inhuman, not when I've seen humans working so warmly with it.
4) One of my favourite quotes about the inadequacy of language (and how we've got to keep trying anyway) is from Douglas Adams: "Let us think the unthinkable, let us do the undoable, let us prepare to grapple with the ineffable itself, and see if we may not eff it after all."
1. Absolutely. I think Arendt acknowledges as much, although perhaps (inadvertently) giving the impression that is unimportant. I believe in the possibility of mystical apprehensions of truth as well as in the eloquence of silence.
2. Agreed as well, and in the same spirit. My sense, certainly based on limited experience, is that fewer of us are trying to keep whacking away, unwilling or unable to renew the well of words.
3. No argument. I think math can convey beauty or an experience of beauty. I suppose Arendt feared that even higher math that sought to describe the world in 20th century physics was increasingly inaccessible to even the most talented individuals. Today, it would seem that certain aspects of the mathematical model of the world or of the technological structures that surround us escape any one person's full understanding. I suppose she might say that the lovely moment that transpires between your son and your husband is something of a different order.
I had some similar thoughts while reading and they felt like arguments to me, but we needn't argue about whether they are or aren't :) I agree words have serious limitations and I'd even go so far as to say they're really not that great. Especially if we're talking about direct communication between human beings, fixating on which words are being used and their precise meanings is often more of a barrier to understanding than an aid. This is based on my personal experience doing exactly that. When I'm talking with people I care about - which incidentally is the use of words that matters most to me in my life - it doesn't seem to matter what specifically is being said as much as what is felt. I think the strongest understanding of relation to others comes from the physical and emotional experience of being in their presence, from their actions both brutal and sensitive, with occasional support from the words they utter.
A (perhaps?) related note is the way that the bureaucratisation of life has led to myriad cottage consulting industries, which are defined often by having access to both a specialised language and specialised methods that use that language. That is, consultants employ algorithms.
Consider how even a coffee shop needs lawyers, consultants, external accountants, supply chain auditors, HR professionals, and SEO optimisers to survive. The professional managerial class has been complicit in forming the algorithmic society by casting high-value knowledge labour as an essentially algorithmic or mathematical process. We only need look at how many high-value firms are essentially just disembodied informational networks: Accounting firms, financial services, banks and so on.
Great post. I look forward to the future installments! The ceding of language for mathematical formulas and black-box algorithmic logic reminded me of what seems a sister issue: the loss of "deep literacy," or a culture that primarily communicates ideas through writing. The author of the piece linked below views deep literacy as a prereq for democracy, which seems in line to Arendt's arguments about literacy and politics.
" It is that our lives are increasingly structured by systems and devices which operate according to mathematical, statistical, and algorithmic logic, a logic that often escapes the understanding not only of those whose lives are subject to the systems and devices but even that of those who create and deploy the systems and devices. " This I think is key. In the past all human inventions could be explained and their theory of operation understood. We knew how and why they worked. With deep learning and large language models even their creators struggle to understand how they're working - how they're coming up with particular answers to questions, etc. What happens when these algorithmic learning models start creating algorithms of their own? This is referred to as "meta-learning" and it's already starting to happen. Meta learning will further separate us from being able to understand our creations.
I was reading Illich's "Medical Nemesis" the other day, and I came across this line which dovetails nicely with this post:
"The religious preference given to scientific language over the language of the layman is one of the major bulwarks of professional privilege. The imposition of this specialized language upon political discourse about medicine easily voids it of effectiveness." (Part IV, ch. 7, "Political Countermeasures: The Scientific Organization--of Life")
What I have to add is most likely not so relevant. As I am reading the newsletter in my email, (at work), the thing that springs to mind immediately is the ludicrous extent to which I'm copied in on messages that are not very relevant to me. Usually, I'll reply to someone who is relevant to my concern, if I have one, but then that creates a separate thread that seems to confuse other people. It's almost as if digital communication has created one of those situations where each branch in the communication stream creates a separate universe. It's becoming increasingly difficult to hold all these worlds in mind, (or in the body, as it sometimes seems). Email has created this crowd of confusion for me, like everyone is standing in my office speaking at once. I used to try to get to every email within two weeks. That time frame is lengthening, and I've been telling people, like our piano tuner who likes to email his invoices, that if they want a timely response, (or payment), they need to send me an actual letter or an actual invoice, through the mail, (like a normal human being - though I do not say that part out loud). That might not have anything to do with the points Arendt or Steiner are making, but it does strike me how estranged from effective communication we have become through communicating electronically. My intuition is this academic office ran much more smoothly and convivially before computers were installed. Eva Brann, (who I have met, though she is a tutor on our other campus), has an essay in the "Minutes of the Lead Pencil Club," where she asks "Cui bono?" in relation to the internet. Who benefits? While there have been some benefits from digital technology, I have yet to see a sound argument for the degree to which digital technology has been adopted.
I'm following this series and the great comments with interest. I tend to join with comments here that regard language as not all of our understanding. The Romantics took a line on sensibility, quote: referring to the painter John Constable's 'Study of Clouds' 1821, ...."He saw skies as fundamental to the development of his large-scale landscape paintings in which it formed the 'chief organ of sentiment'."
I guess because we are embodied with it, we sense both much we don't understand, and much that we understand without words?
Great to touch base again with Arendt and Auden and Steiner. Steiner was always stimulating when I was young. For me 'Science' has a way of generating paradox, and trying to mechanise our logical thinking in order to demonstrate our grasp of reality could be running into limits, mundane as well as perhaps fundamental.
Regarding those limits, a substack writer based in Yorkshire UK, 'Flat Caps and Fatalism', had already read this your latest. and referred us all to it. I left a comment with him which included ..."I have written a review of the book, if I can get round to making it available. It is quite some loom we are tending now." I was referring to Jeremy Naydler's 'In The Shadow of the Machine: The Prehistory of the Computer and the Evolution of Consciousness' https://www.goodreads.com/en book/show/42289015".
My very best hopes for recovery for all those on the Stroke Ward!
Thank you for the good wishes and this stimulating comment. It reminds me of Pasteur: “Chance favors only the mind that is prepared.” There’s a beautiful dance there between recovery and discovery, to the beat of that ever-hopeful value of sensibility (Clouds! Skies! I love that!) the Romantics defended and exalted both in the world and in “ta mathemathike,” (things known): that inward eye that is the bliss of solitude.
I zeroed in on the same passage highlighted by Phil. It seems this diagnosis is consistent with the Habermasian distinction between system and lifeworld and with his theme of the colonization and technicization of the lifeworld by system and its de-linguistified steering media.
Thank you for Pasteur and thoughts; yes, 'recovery and discovery' ... perhaps there is a 'conspiracy for miracles' to be inserted into time? I have had a few minor ones of my own recently.
I have returned now and again to your writings since I first discovered them about a year ago. As a new member let me preface this comment with a very brief introduction. I wouldn't quite call myself a scientist, but I am a humble student of those sciences which are sometimes called "natural". As such I confess to practising the dark arts referred to here, and in this article in particular, which are mathematics and programming.
I can certainly see the modern divergence of mathematics from colloquial language. I agree that talking about certain mathematical or scientific concepts is not "meaningful" in the sense that we cannot relate those concepts to our lived or even imagined experience.
I would like to suggest that part of the confusing nature of mathematics arises from conflation of "mathematics the language" and "mathematics the model". In his book "The Road to Reality", Penrose introduces, in addition to the physical and mental worlds, a world of pure, "Platonic" mathematical objects, or equivalently a world of objective mathematical truths. To paraphrase his reasoning: while ∞ is a symbol, the concept of infinity is a mathematical "entity". These entities are what I would call "models". We have known about and used these entities for a long time (square root of 2, imaginary number, etc.) And yes, some of them cannot be "talked about" except using the language of mathematics. But they do not belong to the "real", physical world. It seems to me that the domain of concepts expressible in natural languages has not shrunk, so much as the domain and use of mathematical models has greatly expanded.
Returning to the social issues, it seems that a different conflation is taking place. The "technologists" see the real world and the world of mathematical models as one and the same, hence the regime of Technique and the structuring of our lives "by systems and devices which operate according to mathematical, statistical, and algorithmic logic". I read this as: structuring of our lives as if they were mathematical models of our lives. While mathematics (the language) can be used to describe some aspects of the real world, its true power comes from the ability to describe entities which exist only in the world of models. Conversely, there are some aspects of the real world, and certainly many of the imagined world, that are only expressible in natural language (or music? art? or not at all?).
I apologize for the length of this comment. One last thought. I don't see GPT-3 as a "triumph of mathematical representations over language", however its success does seem to suggest that the computational complexity involved in using language is lower than previously thought, and this does have ramifications. However, I would not call what ChatGPT is doing "speaking". Yes, this is pedantic, but speech, to me, implies agency, intent. GPT-3 can only "produce text" that follows the rules of natural language and is aware of contextual cues. As a technical aside, there is little mathematics involved: mostly the training data is simply encoded in a huge network of probabilities. The reason they get away with calling this "intelligent" is because that happens to be the best model we currently have for how our brains process certain data. Beware the marketing.
“ In his "Philosophical Investigations" Wittgenstein demonstrates how language only acquires meaning through being embedded in a "form of life".
Understanding language necessarily involves reacting in what the relevant language community (great or small) regards as an appropriate manner.”
This is a beautiful way to view Arendt’s concern (we might almost say love or affection) for free political speech: private words made public when uttered, on the one hand, and on the other, words that can only “mean”—even privately—within repeated as-sociations, in evolving, algorithmic neural networks that say “yes” to “tree.”
As I write this from my bed in the stroke ward of a very advanced hospital, I think your speculations might have gone too far, not because, as one might imagine, I think the math and science saved me, but because they didn’t.
Contemplating neuroplasticity, my work as a high school English teacher and the potential for my recovering my feet, my balance, and my life, I am sensing a sublime calculus in network math that neither Postman nor Arendt nor Steiner knew or considered.
I’ll have more to say about this later, but the gap you and they sense between speech and ineffable mathematics might be summarized as the gap between the simulated model and the lived experience.
And one of those models is the 20th century, distinctive mindset. Arendt’s distinction, for example, between public and private establishes a foundation for her ongoing argument in The Human Condition. But read the last 30 or 40 pages of that book with a literary ear, and you can hear her feeling loss. It’s downright poetic. Thus, too, Steiner.
Our new world is a networking world, one where mathematical models that operate in neural networks help to prepare us for the strokes to come: climate blasts of major seaports, solar flares, that knock out electrical grids, all kinds of losses of infrastructures and institutions that we created on the foundation of critical distinctions like “public and private.”
Those distinctions can be modeled so that we can reinvent them, and the institutions on which we built our way of being together.
Mathematical modeling, especially with networks and network math, helps us to prepare to heal. There is much more to say on this, and I understand and apologize for the abstraction of that position. But my literal physical position, flat on my back and served by the neural networks that allow me to dictate this post response into my phone, convince me that our artistic heritage (think, say, of Russian formalist Viktor Šklovsky, arguing that by making language strange, poetry helps us live again), including the flow or apparent flow of free speech, has much still to offer us as we render that flow in mathematical models.
We are seeing a convergence, not a loss.
Paul, I want to wish you the very best on your recovery. Your comment led me to revisit this conversation between Nils Gilman and Benjamin Bratton from last year: https://www.noemamag.com/a-new-philosophy-of-planetary-computation/
I found it fascinating and thought you might too.
Thank you, Abby, for your good wishes as well as this remarkable conversation link. It answers to my intuitions, philosophical and otherwise, and does a fine job of suggesting a framework for the kind of thesis my comment was clawing to articulate. Do you write a blog? I’d love to follow it.
I don't, actually, but it was very kind of you to ask. Glad you enjoyed the article.
Paul, Thank you for this. I'm sorry to hear of the stroke, but do hope your recovery is progressing. There's much too think about in what you've written and given your response to Abby below, I'd like to take the exchange between Gilman and Bratton into consideration as well. I confess I've had that exchange and several essays that have appeared on Noema bookmarked for some time. I hope to maintain a posture of critical openness to more hopeful outcomes, to be prepared to be surprised as Illich might have put it.
Not arguments, just a few thoughts that occurred to me while reading this:
1) My experience of my own brain is that there is a lot happening in there that is also non-verbal and unseen, so I'm not quite sold on the idea that words deserve quite so much primacy. And between humans, too, there's many a slip. We were never communicating perfectly with words, even when that was all we had.
2) As a writer of fiction, I feel keenly aware of the limitations of words. There are ways I feel and experiences I've had that words are absolutely inadequate for. I would even say that the most important things (to me) are the hardest to convey in words. Language is a blunt instrument. I love it, and will keep whacking away with it, but there's a reason we need music and paintings and dance as well.
3) One of my favourite things is when I'm on one side of the kitchen table, writing or drawing, and my husband and son (a physicist and engineer, respectively) are on the other side of the table working calculus problems together. This is as cozy a father-son bonding moment as I have ever witnessed, as incomprehensible as their mouth-noises are to me. So I don't quite buy that the language of math is only ever cold and inhuman, not when I've seen humans working so warmly with it.
4) One of my favourite quotes about the inadequacy of language (and how we've got to keep trying anyway) is from Douglas Adams: "Let us think the unthinkable, let us do the undoable, let us prepare to grapple with the ineffable itself, and see if we may not eff it after all."
Thanks for these, Rachel.
1. Absolutely. I think Arendt acknowledges as much, although perhaps (inadvertently) giving the impression that is unimportant. I believe in the possibility of mystical apprehensions of truth as well as in the eloquence of silence.
2. Agreed as well, and in the same spirit. My sense, certainly based on limited experience, is that fewer of us are trying to keep whacking away, unwilling or unable to renew the well of words.
3. No argument. I think math can convey beauty or an experience of beauty. I suppose Arendt feared that even higher math that sought to describe the world in 20th century physics was increasingly inaccessible to even the most talented individuals. Today, it would seem that certain aspects of the mathematical model of the world or of the technological structures that surround us escape any one person's full understanding. I suppose she might say that the lovely moment that transpires between your son and your husband is something of a different order.
4. lol
I had some similar thoughts while reading and they felt like arguments to me, but we needn't argue about whether they are or aren't :) I agree words have serious limitations and I'd even go so far as to say they're really not that great. Especially if we're talking about direct communication between human beings, fixating on which words are being used and their precise meanings is often more of a barrier to understanding than an aid. This is based on my personal experience doing exactly that. When I'm talking with people I care about - which incidentally is the use of words that matters most to me in my life - it doesn't seem to matter what specifically is being said as much as what is felt. I think the strongest understanding of relation to others comes from the physical and emotional experience of being in their presence, from their actions both brutal and sensitive, with occasional support from the words they utter.
A (perhaps?) related note is the way that the bureaucratisation of life has led to myriad cottage consulting industries, which are defined often by having access to both a specialised language and specialised methods that use that language. That is, consultants employ algorithms.
Consider how even a coffee shop needs lawyers, consultants, external accountants, supply chain auditors, HR professionals, and SEO optimisers to survive. The professional managerial class has been complicit in forming the algorithmic society by casting high-value knowledge labour as an essentially algorithmic or mathematical process. We only need look at how many high-value firms are essentially just disembodied informational networks: Accounting firms, financial services, banks and so on.
Great post. I look forward to the future installments! The ceding of language for mathematical formulas and black-box algorithmic logic reminded me of what seems a sister issue: the loss of "deep literacy," or a culture that primarily communicates ideas through writing. The author of the piece linked below views deep literacy as a prereq for democracy, which seems in line to Arendt's arguments about literacy and politics.
https://nationalaffairs.com/publications/detail/the-erosion-of-deep-literacy
" It is that our lives are increasingly structured by systems and devices which operate according to mathematical, statistical, and algorithmic logic, a logic that often escapes the understanding not only of those whose lives are subject to the systems and devices but even that of those who create and deploy the systems and devices. " This I think is key. In the past all human inventions could be explained and their theory of operation understood. We knew how and why they worked. With deep learning and large language models even their creators struggle to understand how they're working - how they're coming up with particular answers to questions, etc. What happens when these algorithmic learning models start creating algorithms of their own? This is referred to as "meta-learning" and it's already starting to happen. Meta learning will further separate us from being able to understand our creations.
I was reading Illich's "Medical Nemesis" the other day, and I came across this line which dovetails nicely with this post:
"The religious preference given to scientific language over the language of the layman is one of the major bulwarks of professional privilege. The imposition of this specialized language upon political discourse about medicine easily voids it of effectiveness." (Part IV, ch. 7, "Political Countermeasures: The Scientific Organization--of Life")
What I have to add is most likely not so relevant. As I am reading the newsletter in my email, (at work), the thing that springs to mind immediately is the ludicrous extent to which I'm copied in on messages that are not very relevant to me. Usually, I'll reply to someone who is relevant to my concern, if I have one, but then that creates a separate thread that seems to confuse other people. It's almost as if digital communication has created one of those situations where each branch in the communication stream creates a separate universe. It's becoming increasingly difficult to hold all these worlds in mind, (or in the body, as it sometimes seems). Email has created this crowd of confusion for me, like everyone is standing in my office speaking at once. I used to try to get to every email within two weeks. That time frame is lengthening, and I've been telling people, like our piano tuner who likes to email his invoices, that if they want a timely response, (or payment), they need to send me an actual letter or an actual invoice, through the mail, (like a normal human being - though I do not say that part out loud). That might not have anything to do with the points Arendt or Steiner are making, but it does strike me how estranged from effective communication we have become through communicating electronically. My intuition is this academic office ran much more smoothly and convivially before computers were installed. Eva Brann, (who I have met, though she is a tutor on our other campus), has an essay in the "Minutes of the Lead Pencil Club," where she asks "Cui bono?" in relation to the internet. Who benefits? While there have been some benefits from digital technology, I have yet to see a sound argument for the degree to which digital technology has been adopted.
I'm following this series and the great comments with interest. I tend to join with comments here that regard language as not all of our understanding. The Romantics took a line on sensibility, quote: referring to the painter John Constable's 'Study of Clouds' 1821, ...."He saw skies as fundamental to the development of his large-scale landscape paintings in which it formed the 'chief organ of sentiment'."
I guess because we are embodied with it, we sense both much we don't understand, and much that we understand without words?
Great to touch base again with Arendt and Auden and Steiner. Steiner was always stimulating when I was young. For me 'Science' has a way of generating paradox, and trying to mechanise our logical thinking in order to demonstrate our grasp of reality could be running into limits, mundane as well as perhaps fundamental.
Regarding those limits, a substack writer based in Yorkshire UK, 'Flat Caps and Fatalism', had already read this your latest. and referred us all to it. I left a comment with him which included ..."I have written a review of the book, if I can get round to making it available. It is quite some loom we are tending now." I was referring to Jeremy Naydler's 'In The Shadow of the Machine: The Prehistory of the Computer and the Evolution of Consciousness' https://www.goodreads.com/en book/show/42289015".
My very best hopes for recovery for all those on the Stroke Ward!
Thank you for the good wishes and this stimulating comment. It reminds me of Pasteur: “Chance favors only the mind that is prepared.” There’s a beautiful dance there between recovery and discovery, to the beat of that ever-hopeful value of sensibility (Clouds! Skies! I love that!) the Romantics defended and exalted both in the world and in “ta mathemathike,” (things known): that inward eye that is the bliss of solitude.
I zeroed in on the same passage highlighted by Phil. It seems this diagnosis is consistent with the Habermasian distinction between system and lifeworld and with his theme of the colonization and technicization of the lifeworld by system and its de-linguistified steering media.
Thank you for Pasteur and thoughts; yes, 'recovery and discovery' ... perhaps there is a 'conspiracy for miracles' to be inserted into time? I have had a few minor ones of my own recently.
A view of the Constable painting was available yesterday https://twitter.com/ahistoryinart/status/1622886400877002755
Best clouds!
Greetings,
I have returned now and again to your writings since I first discovered them about a year ago. As a new member let me preface this comment with a very brief introduction. I wouldn't quite call myself a scientist, but I am a humble student of those sciences which are sometimes called "natural". As such I confess to practising the dark arts referred to here, and in this article in particular, which are mathematics and programming.
I can certainly see the modern divergence of mathematics from colloquial language. I agree that talking about certain mathematical or scientific concepts is not "meaningful" in the sense that we cannot relate those concepts to our lived or even imagined experience.
I would like to suggest that part of the confusing nature of mathematics arises from conflation of "mathematics the language" and "mathematics the model". In his book "The Road to Reality", Penrose introduces, in addition to the physical and mental worlds, a world of pure, "Platonic" mathematical objects, or equivalently a world of objective mathematical truths. To paraphrase his reasoning: while ∞ is a symbol, the concept of infinity is a mathematical "entity". These entities are what I would call "models". We have known about and used these entities for a long time (square root of 2, imaginary number, etc.) And yes, some of them cannot be "talked about" except using the language of mathematics. But they do not belong to the "real", physical world. It seems to me that the domain of concepts expressible in natural languages has not shrunk, so much as the domain and use of mathematical models has greatly expanded.
Returning to the social issues, it seems that a different conflation is taking place. The "technologists" see the real world and the world of mathematical models as one and the same, hence the regime of Technique and the structuring of our lives "by systems and devices which operate according to mathematical, statistical, and algorithmic logic". I read this as: structuring of our lives as if they were mathematical models of our lives. While mathematics (the language) can be used to describe some aspects of the real world, its true power comes from the ability to describe entities which exist only in the world of models. Conversely, there are some aspects of the real world, and certainly many of the imagined world, that are only expressible in natural language (or music? art? or not at all?).
I apologize for the length of this comment. One last thought. I don't see GPT-3 as a "triumph of mathematical representations over language", however its success does seem to suggest that the computational complexity involved in using language is lower than previously thought, and this does have ramifications. However, I would not call what ChatGPT is doing "speaking". Yes, this is pedantic, but speech, to me, implies agency, intent. GPT-3 can only "produce text" that follows the rules of natural language and is aware of contextual cues. As a technical aside, there is little mathematics involved: mostly the training data is simply encoded in a huge network of probabilities. The reason they get away with calling this "intelligent" is because that happens to be the best model we currently have for how our brains process certain data. Beware the marketing.
“ In his "Philosophical Investigations" Wittgenstein demonstrates how language only acquires meaning through being embedded in a "form of life".
Understanding language necessarily involves reacting in what the relevant language community (great or small) regards as an appropriate manner.”
This is a beautiful way to view Arendt’s concern (we might almost say love or affection) for free political speech: private words made public when uttered, on the one hand, and on the other, words that can only “mean”—even privately—within repeated as-sociations, in evolving, algorithmic neural networks that say “yes” to “tree.”