Tim and I read McQuillan, Dan. Data Science as Machinic Neoplatonism. Philos. Technol. 31, 253–272 (2018). https://doi.org/10.1007/s13347-017-0273-3
Key themes discussed:
- The critique of vulgar vs. authentic Platonism in data science
- Process-relational philosophy and constructive realism
- Evolutionary biology’s relevance to understanding AI/machine learning
- The political and ethical dangers of algorithmic decision-making
- The relationship between technology, magic, and science
- Alienation in digital technologies and the concept of “digital twins”
- The “Platonic representation hypothesis” in machine learning
- Participatory vs. onlooker consciousness in science
Transcript:
Matt Segall: Hey Tim.
Timothy Jackson: Hey, man, how you doing.
Matt Segall: Cool sweater—are those sea slugs?
Timothy Jackson: Yeah. Nudibranchs.
Matt Segall: Nice.
Timothy Jackson: Naked gills.
Matt Segall: So you are underslept and overread?
Timothy Jackson: Oh, yeah, big time. I mean, that’s pretty much part of the course. My daughter’s not sleeping well, my dog’s not sleeping well, so we’re not sleeping well. Even when they are sleeping well, I guess one is anticipating being woken up, and then you come to overvalue sleep, and of course overvaluing sleep is a guaranteed way not to get enough—the royal road to insomnia.
So yeah, but you know, all good. But yeah, overread as well. Like I was telling you yesterday, when I should have been revisiting Mcquillan for today’s discussion, I was reading—rereading Anderson, P.W. Anderson on “More is Different” from my cybernetics reading group on Friday. And that’s just such a generative paper. It’s so short, it’s like four pages, but I have like more than ten pages of notes from yesterday. I just went down kind of every rabbit hole in there.
Yeah, no, it was good fun, but didn’t help me prepare for this chat.
Matt Segall: Yeah, no. I aspire to be able one day to publish a four-page article that will just offer an open and shut case about whatever the issue might be. But good reminder that brevity and concision are actually way more potent than… I can do it. Yeah, real point.
Timothy Jackson: It’s four pages in the old science format, which I really don’t actually like reading on my remarkable, because it’s two columns, right? So you know, in many other journals it would be an eight or more page paper, probably. So it’s not tiny, but he really packs it full of things. But a lot of stuff is left—there are a lot of implications. So one is left to unpack a lot of things, but it is, you know, it’s the seminal paper on emergence, really, for people who want a kind of harder physicsy complexity science approach to emergence. It’s still the go-to.
Matt Segall: Yes. Well, we will certainly do a session at some point in the future, as I’m sure we will continue these dialogues for decades.
So yeah, let’s dive in to the Mcquillan article “Data Science as Machinic Neoplatonism.” And I think, as one of us mentioned last time when we were supposed to talk about this paper but barely touched on it, just because Novalis and Simondon are so interesting—you know, I think I said it’s very prescient. It was published in 2015, or maybe even it was written in 2014 or no, received 2015. Wow! So we’re going back almost a decade.
Timothy Jackson: Drafted, I guess we could say drafted in 2014, and so…
Matt Segall: Very prescient, because I think, you know, people were aware of the algorithms that were driving social media. And that was part of the public conversation to some degree, especially since the Trump election, when it really seemed like, “Oh, wow! These are mind control devices. And this is a national security threat,” and, like, you know, that really became part of the public consciousness back then. But you know, now everyone’s obsessed with large language models and artificial intelligence.
And I think we’ve in some ways forgotten about the extent to which the data that all of these systems are being trained on might contain biases that need to be noticed, and that we can’t just treat these technologies as a kind of neutral intelligence, because it’s trained on us and our cultural discourse, and so is fully tainted by any bias that it gets fed on the data it’s trained on.
And so yeah, glad we’re spending some time with this one. Before I hand it over to you for some initial impressions, as you might expect, I’m going to need to stand for Plato and my pointless boys a little bit here, and ladies, you know, Hypatia and Conway. And this isn’t just a gentleman’s club Platonism.
But I think maybe what I will end up wanting to do is articulate a countercultural Platonism or Neoplatonism along the lines of the countercultural data science that Mcquillan wants to inaugurate. So we’ll come to that. But yeah, maybe I’ll just let you offer your initial impressions on your sense of this paper.
Timothy Jackson: That was almost a very interesting parapraxis—little slip there. My impressions.
Yeah, no, I figured that’s where we were gonna go in terms of a discussion of so-called Neoplatonism and its deployment here. And yeah, I mean my concerns, worries… Am I worried? Am I just irked by what I see as a kind of alliance in a contemporary—you know, it’s not just data science, it’s kind of becoming a mainstream paradigm, I think, in science in general—an alliance between something coming out of second-order cybernetics, like a constructivist position coming out of…
Matt Segall: Sure.
Timothy Jackson: Yeah, Heinz von Foerster being like a very key example, and key influence, and this kind of Platonic realism thing. So a lot of name-dropping of Plato or Platonism.
And I’m yeah, I have my, as you know, I have my concerns about the way those things sit together, and I think that’s where we’ll go for the meat of this conversation, I suspect. Yeah, and I know you’ll want to defend more nuanced usage of Plato’s name, which I think is perfectly adequate, like important for you to do that. And I also think it’s interesting that in many ways the kinds of positions that you and I are always trying to articulate and working on together, despite our little bit of difference that we have between one another, but our sort of shared position—some sort of constructive, you know, process-relational, constructive realism kind of thing—is in many ways an alliance between a kind of constructivist view and a Platonic view. You know, there’s a relationship there that we’re trying to chart.
What I’m concerned about is a kind of oscillation between the two, like a kind of pinging back and forth without a capacity to occupy that very, very slippery middle ground. And then, I think, you know, we’ll get to the way in which it seems like an oscillation as well across some sort of spectrum, because, you know, constructivism can seem kind of like anti-realism, and then you’ve got this hard realism on the other side, in a sense, right? But they’re also kind of a collapse onto the same position in a really interesting way. And so I think we get to that.
But I think, yeah, there’s so much in the Mcquillan paper other than that. And you know, I just looked like ten minutes ago, and I’ve got lots of notes in the document. But they’re not all going to be in my working memory. Maybe we can work through some of the stuff directly from the document, or perhaps you have some notes that will guide us.
You know, on a couple of things that you just said: Yes, you know, it’s prescient, and I just judge—I don’t know Dan Mcquillan’s work well. But as we discussed last time, I came across him in association with some of the podcasting that he did for the release of his 2022 book Resisting AI, and I guess, funnily enough, he was talking, I think, in some of those interviews about how he’d written this book resisting AI, and then the LLMs kind of went public and took off, you know, in a big way, and he probably would have said something slightly different, but maybe even sharper, if that had happened, you know, during the writing or prior to the writing. But it’s obviously something that he’d been working on for a long time, because even this is not his first paper. He’s got references in there to, I mean, around that same time, like 2014, 2015, to papers in this subject area. And so it’s definitely prescient in a lot of ways. And I think some of the problematics that he points to have just got worse.
Including, as we’ve already hinted, the alliance between a certain take on a Platonist metaphysics and data science. And for Mcquillan, that’s kind of a tacit thing, right? You know, like he thinks that data scientists aren’t explicitly framing their work as Neoplatonist or Platonic, or Platonistic, or whatever the term we want to use is, but that there’s this, you know, he thinks potentially dangerous tacit alliance which we’ll get into. But it’s become more explicit, I think, you know.
Matt Segall: Yeah. Yeah. Pseudo-platonic, how about that?
Timothy Jackson: Okay. Pseudo-platonic.
Matt Segall: So just—I mean just to tease where I want to go—I think what he’s criticizing and referring to as Neoplatonism is actually more like neo-gnosticism.
Timothy Jackson: Okay. Yeah.
Matt Segall: And you know, because then we can see how like Plotinus has very significant criticism of the Gnostics for making matter just evil and the body just evil. And you know, I’m going to talk about Iamblichus and the importance of materiality as a sort of creating a home for the gods, that the material world is not fallen and irredeemable. Far from it, we need it to actually access the divine through ritual and through, you know, material, semiotic, material discursive practices. And so there’s so much in the Neoplatonist—if you actually do a closer reading of the texts—that is saying exactly what Mcquillan is arguing for in terms of a countercultural data science.
Timothy Jackson: Could we call it—yeah, sorry. Could we call it like vulgar Platonism in a way, like… And similarly, we talk sometimes about this term that’s going to come up in this—like naive Kantianism, or something like that, or vulgar Kantianism, or something like that. And there’s this, again, very close relationship between what I would mean by that and this sort of vulgar Platonism, I think.
And that’s part of how this constructivist and Platonistic or vulgar Platonist position come into relationship. Now, I think that the Iamblichus reference is really, really nice, and that just reminds me of a kind of—you know, there’s a vulgar Buddhism which has a kind of almost two-worlds aspect to it as well, right? You know, that we’re trying to escape Samsara, you know, get off the wheel of suffering, you know, and if we just meditate enough, we can do it in like one lifetime, instead of having to slowly climb the ranks of beings towards some sort of more enlightened position over many, many reincarnations. But then there’s the counter-movement, which is the seeing the sacred in Samsara, which is, you know, again, like this is all actually you’ve got, and it’s like all you’re going to get, and the death and rebirth cycle, then, is, in fact, a psychic movement, a psychological process which enables you to become, you know, more enlightened, but really to again see the sacred in the fallen world, or to mix traditions, but in Samsara. And that’s where I like to go with my notion of prosaic realism.
Matt Segall: I mean, before we—I definitely have a lot to say about what gets rendered as Neoplatonism in this article, but I also fully and completely am aligned and in agreement with what Mcquillan is trying to resist and the way he’s pointing out the dangerous potentials of algorithmic violence and discrimination and the evasion of due process. And here we are in 2025, and you know, Elon Musk and Peter Thiel and others have installed their proprietary AI systems right into the heart of the U.S. government. These algorithms are being used by Immigration and Customs Enforcement to find, arrest, and deport people who have no opportunity for due process, and no doubt will soon be used against U.S. citizens. And you know, all around the world this is happening. And so the ship has sailed—it was already being used in these ways, you know, a decade ago, when Mcquillan was writing this. But now it’s everywhere, right? And so this is kind of urgent. This is an emergency situation. And there’s a real activist bent to what Mcquillan is writing here, and you know, he wants to be more than just rhetorical in his critiques of data science and offer some examples, you know, at the very end he offers some examples of the use of data science to actually foster community engagement and forms of resistance to, you know, the more top-down military and state and corporate-based uses of this technology.
And so, you know, he’s giving us a lot of rhetorical tools also, though, you know, so that we don’t fall into the trap of the PR around this stuff, that these are just neutral technologies, and that there is such a thing as general intelligence that could be artificially realized, which backgrounds all of the cultural assumptions and the paradigmatic assumptions, you know, that are built into these technologies. I love his whole point that he begins with, you know, drawing on Kuhn, that this is a whole paradigm and an organizing idea. It’s not just a methodology.
And the importance of standpoint epistemology and situating science culturally and historically—this is really important. I’m glad that Sandra Harding’s work is great. I’m glad that Donna Haraway was also mentioned, who’s like a white and indigenous feminist philosopher of science, because she argues for kind of situated objectivity, so she’s not chucking some conception of scientific objectivity—she actually wants to preserve situated objectivity, but she’s resistant to, you know, the idea that, say, the body biologically speaking—she’s a biologist, too, right? She doesn’t want the body to just become a blank slate for us to technologically inscribe with whatever we want. It’s like, no, the biological world is constructed in an evolutionary sense, but still is going to resist any forms of cultural or conceptual—you know, it’s the body. The body is going to resist our conceptualizations of it. So we can’t just write on it whatever we want. So that’s…
Timothy Jackson: Which, you know, the sophisticated form of constructivism or realism, where there is this kind of layering of, you know, enabling constraints that are constructed over evolutionary time. And in a lot of ways, the basic—you know, we’re harping on this all the time—but you know, in a lot of ways, the basic principles are not different from kind of social constructivism. There are very different mechanisms involved.
But you know, there are these processes happening over different temporal scales, and that leads to both enablement but constraint in general, which is part of why, you know, a body will resist certain kinds of facile attempts to rewrite it at some or reconceive it at some conceptual level, because biology is a real thing. Physics is real. There’s a real deep evolutionary history associated with the properties and functions that a body has.
I mean, just to riff really quickly on a couple of—I mean, I love Donna Haraway, by the way, so I like that. And of course he makes a lot—he does a lot of work with Karen Barad in this essay as well, right? But just to your point, right at the beginning, about the biases and the fact that, you know, the training data is essentially us and supplied by, annotated by, you know, humans. So therefore, containing biases that we have, you know, we’re preloading any of the results with priors, basically axioms, almost.
I mean, I think there’s just a garbage in, garbage out principle, obviously, with these things. And I actually think that’s also part of a pushback against a kind of facile criticism even of LLMs, you know, like LLMs—you can have a conversation with an LLM that’s going to be basically as good as your prompt in some sense, you know. If you feed it a bunch of crap, it’ll turn a lot of crap out, but if you provide it with very well-curated content, and it’ll rapidly degenerate, and so you have to keep, you know, bringing it back on task. You have to have a very critical relationship with these things, and indeed you should encourage them to have a critical relationship with you, because one of the big dangers here is a kind of echo chamber amplification of whatever our personal fantasies are. But if you have that kind of reciprocally critical relationship—obviously these are not real agents, but that you would have with another agent, with an interlocutor, that you were thinking constructively with—then, you know, you can have a constructive dialogue with these things, which is a very different issue from many of the issues that Mcquillan is highlighting.
And he’s also sharpened a lot of the criticisms that we have in this paper over time. So he does at the end of this paper talk about the possibility of a machine learning for the people, right? But I think he’s given up on that. In fact, I gather his book Resisting AI was originally titled when he first had the proposal and got the contract for writing it “AI for Good.” So he was actually wanting to write a book about machine learning for the people, basically. So what are the ways that we can use this to increase human flourishing, whatever, and ethically, etc. And the more he dug into it, the more research he did, the more he just was disillusioned with that goal. And the more he—I think he basically says that these are just an instantiation and intensification of neoliberal principles.
Matt Segall: Dope.
Timothy Jackson: His concerns have got worse, and with all the preemptive stuff that he talks about here, and the sort of algorithmic force. But there’s a lot of stuff about preemption in the legal context, and you know, the testimony of the model being having, you know, more weight than the testimony of a human witness, or indeed the person on trial, or whatever. It really reminds me—I guess it’s a facile comparison to draw—a bit of Minority Report. And then the way that these self-fulfilling prophecies enter into the ecosystem. And this reminds me of Dune as well, and the Atreides and their capacity to predict the future. But, in fact, in the later Dune books it becomes revealed that they’re not so much predicting the future as manifesting it, because basically they are such powerful loci of agency that their vision for the future is essentially what’s bringing that particular future into reality. And I think that’s a real concern here. That’s the machinic aspect. I don’t think Deleuze and Guattari are actually mentioned in this paper, but obviously the conception of machinism that’s being deployed here is deeply, you know, originally Guattarian, and developed by Deleuze and Guattari, and he talks at a certain point also about—so they’re machinic in the sense that they’re just acting directly in the world.
So they are maybe not loci of agency if we want, but they are loci of force—the algorithm, of course, that he talks about, and he talks about them as machines that produce or mold subjectivity, right? And that’s a very Foucauldian position. And he references Foucault. But that’s exactly Guattari’s position as well, right? Processes of subjectivation, you know.
Matt Segall: Yep, yep. And you know, on page—the second page of this article, when Mcquillan is critiquing the sort of positivist assumptions that go into a lot of data science, the naive empiricism, you know, he says, “Reality is not directly accessible to us as facts that can be recorded by devices.” And he says, “Meaning is not an object of sensory perception.”
Timothy Jackson: Hmm.
Matt Segall: And so when I read that I was like, this is basically Plato’s epistemology. You know. And so he’s kind of—this is an example of how I think very often there’s a lack of appreciation for how much of our capacity for critical reason was actually, you know, inaugurated in the Platonic dialogues. And so you know, this—avoiding the myth of neutral facts and recognizing that meaning is not just read off of sensory perception—that’s the whole point of the distinction between the sensible and the intelligible, as I read it.
Timothy Jackson: Yeah, yeah.
Matt Segall: You know, in Plato. And so for Plato, the sensible realm is an image of the intelligible.
Timothy Jackson: Yeah, I mean, I don’t know. It’s the whole—I mean, maybe it’s the proximal point, in some sense, like it’s the primary point. I think that therefore the ontology is—it is problematic. I mean, and not necessarily just in the good sense that Platonic ontology, which doesn’t necessarily correspond directly with that epistemological point—it’s very related to it. There’s an onto-epistemology here, right? But I think that again, we’ve been through this with the Timaeus, but the way the ontology gets enshrined in a sense, even though it’s problematized imminently within the dialogues, doesn’t necessarily correspond, and I think that—you know, this is also the Kantian point, right? You know, the phenomenal/noumenal distinction is the same point, but the ontology is kind of different. Again, it’s not my ontology, but the ontology is different, and I think it’s just a generic—like he attributes the point to Kuhn. I don’t want to derail you, and you’re making a really, really important point about Plato. He attributes the point to Kuhn. Kuhn gets it as much as anything from Popper as well. You know, Popper says that all observation is theory-laden, and Popper gets it from Einstein. Like, I think Popper directly credits Einstein with, you know, the—what is it?—”The theory makes the observation possible,” or whatever, “determines what we can…”
Matt Segall: Probably got that from Mach, too, right? I mean.
Timothy Jackson: Yeah, yeah, I mean, it’s just a perennial…
Matt Segall: All goes back to—who? I don’t know.
Timothy Jackson: Yeah, well, Plato is your point.
Matt Segall: Well, yeah, yeah. And so I think, right on that same page, too, in the next paragraph, there’s this comparison made between data science and the idea of a hidden mathematical order that’s ontologically superior, you know, to the one available to the senses. And I think that’s actually—again, this is more Gnostic. I think it’s actually the Gnostics who have the two-worlds theory, not the Platonists who are critical of the Gnostics, because in Plato, the mathematical order is not hidden. It’s the harmony and proportion of the visible cosmos.
Timothy Jackson: Yeah.
Matt Segall: It’s the visible cosmos. It’s the motion of the stars and planets that taught us math, Plato says.
Timothy Jackson: We learn it inferred, though, in a sense, like, which I think is appropriate. But isn’t it also explicitly stated that the demiurge is the only one with direct access to the forms in the cosmological scheme, and then also in the Symposium and other places, the pre-incarnate soul has access to the world of forms that the body, the incarnate body, or the incarnate soul doesn’t have access to. And so you have to go through this anamnesic ascension to kind of recover the knowledge that is innate because you got it prior to incarnation. And so when you say it’s Gnostic, just—I agree that the Gnostics make it very, very distinct, but…
Matt Segall: Maybe they’re the first vulgar Platonists. They’re the first ones to…
Matt Segall: because they’re not Platonists. I think that’s yeah.
Matt Segall: The Gnostics maybe are taking Plato and refusing, or they’re unable to hover in the in-between and hold the tension of opposites. Yeah.
Matt Segall: But I think what distinguishes the Platonic tradition from the Gnostic vulgarization of it is precisely that capacity to hover between. And this is the same sort of hovering that Novalis, and the whole romantic tradition, I think, recovers, you know, holding that tension. And so—but you know, I think still, whether or not it’s really Neoplatonic, or it makes sense to refer to it that way—with the point Mcquillan is making, I fully agree with, and this way of making the math superior to the experienced world is very problematic and forcing a mathematical fit.
You know, to the way that the algorithms are designed and the use of the sigmoid function that he explains in a helpful way as an assumption, you know, built into the design of the algorithm, not something we’re discovering in the data, so much, right? And so to force mathematical fits, it just lends itself—because these models end up being so predictive, it’s very tempting to collapse statistical correlations with causal and physical explanations. And so the risk here is that we end up with a kind of science that doesn’t actually explain anything to us in a way that human beings can understand, at least, but that functions to make preemptive decisions about how the world, how the human world, the social world should unfold.
And so there’s a real danger here to science, I think, that it’s hijacked by this conflation of statistical correlation with the sorts of causal explanation that science, I would think, would want to seek.
Timothy Jackson: Yes, but I think that that is a danger that radically predates machine learning. So I think that danger’s just…
Matt Segall: But the power of these machines to do these statistical correlations…
Timothy Jackson: But the power of the formalisms, like, you know, the—obviously mechanical philosophy and Cartesian heritage, blah blah, but in a sense the Laplacian paradigm. The power of the formalisms to do something very similar to that and kind of to confuse a limit—like determinism as an emergent limit with the immanent process at the level of the messy, messy details, because it does give us—it gives us a certain kind of very, very powerful grip, because, you know, nature in the broadest possible sense does exhibit a great deal of regularity.
And so these methods give us a lot of instrumental purchase on the world, and then we confuse that for a—you know, we tacitly, in a sense, reify an ontology associated with or read off the formalism that gives us an instrumental grip. And so I think, you know, actually, this is what Mcquillan’s kind of saying, right? He’s saying that this was basically the ontology of like Copernicus and Galileo, and you know, nature written in the language of mathematics, and all of that. And this is just an intensification of that. So I think, yeah, we’re just making the same point there. It is a danger to science, because it’s just also old school reductionism. Again, you know, like, it’s assuming that the—again, those regimes of reality which are most amenable to description in these formal terms are therefore, you know, ultimate, and everything else is in principle, if not in practice, reducible to them. But we see—and we’ve continuously had this problem, and it’s forced paradigm shifts in biology and other places—that you simply can’t use those formal—it’s not that they’re not useful, or that, you know, models derive from that lineage, which is basically all of the mathematical models that we use, are not useful in biology and elsewhere. But they’re not going to give you predictive purchase on what an organism is going to do next, or indeed where evolution is going or anything like that. And this is why we have, you know, people like Stuart Kauffman, and the notion of unprestateability and adjacent possibility, and all of that. And we, you know, we need a different ontology, basically. And it needs to not be vulgar Platonism, but it needs to be a much more reciprocal participatory relationship between those worlds which are in some sense two limits of the same immanent process. And that’s why, you know, some of the philosophers that we really love are really good at providing—again, you can call it an immanent critique, because they’re still making use of the Platonic operators, but—of trying to really occupy that middle ground and really figure the relationship between the actual and the virtual, or, you know, form and material, or whatever your duality is, a little bit more rigorously.
I can’t resist—like we said, you know, goes back to Plato, the kind of appearance/reality sort of distinction. But I mean, it’s just all over in the pre-Socratics, too, right? I mean, it’s Democritus—like people neglect that the Democritean point about atoms in a void is also that kind of point, you know, like, it’s also a point about, you know, you see all of these qualities and everything, but really, ultimately it’s just atoms in a void. There’s a kind of skepticism about appearances that is embedded there as well. Parmenides’ way of truth, way of opinion, I mean the Vedic tradition, Maya, you know…
Matt Segall: Right, right, I mean the appearance/reality distinction would precede Plato, but the idea that the sensible is an image of the intelligible…
Timothy Jackson: Yeah, that may. Yeah. Okay, yeah.
Matt Segall: I think is where the—you know, that’s the Platonic epistemology gets its—that’s his novelty, I think. And, you know, you mentioned participation. Methexis is straight also, straight out of the Platonic tradition. And so, as is the, you know, the very concept of epistemic injustice, that knowledge and truth should have some necessary ethical vector to them oriented towards the good, right. This, again, is part of the Platonic lineage. And so there’s…
Timothy Jackson: In a sense, again to preach very much to the pulpit here in the sense of like, what is your handle that you’ve been using for the last, I don’t know, ten years, or whatever in your online presence? And Whitehead’s quote “footnotes to Plato.” It’s almost—it’s not saying a lot to say that these things which have dominated the history of Western philosophy are to be found in the Platonic tradition, right? It’s like, but even…
Matt Segall: Newer things like the idea of historicizing. This is the last point I’ll make, maybe about this, and then we can—I think it’d be helpful just to focus on the importance of what Mcquillan is arguing. But the idea of historicizing science, if you recall that Timaeus begins with an account of ancient Atlantis and influence on the Egyptians, and how their understanding of number and geometry is what influenced the Greeks. And so you know, before Timaeus can get going, offering a cosmology, they have to situate their capacity to even know what number and math are about in this historical process, you know, of inheriting an understanding from prior civilizations. It’s also a place-based understanding of knowledge, even, you know. And so there’s a demonstration, I think, in the very structure of the narrative, in the dialogues, that scientific knowledge is culturally mediated, historically situated and place-based. You know, it’s all there.
Timothy Jackson: Yeah, no, I think it’s a point very well taken, and I think it would be very foolish to argue against the over-determining, but in both a positive and a negative way, I think, influence of Platonism. I think we really do need to problematize this implicit and then subsequently reified, maybe in the Gnostics, two-world metaphysics. But meanwhile we need to focus on that middle ground which was also staked out in that position, you know.
Matt Segall: So my Platonic or Neoplatonic critique of data science would be exactly Mcquillan’s critique. It’s this idea of insight through opacity and this dichotomy that ends up opening up between this rift between human Logos, human reason and machine logistics, you know, that Mcquillan highlights. If we can’t understand it as human beings, then it’s not truth, it’s not science. It’s not reason. You know, that’s a—and I feel like this is—one can marshal the Platonic tradition to make precisely that point. Insight through opacity is not a thing.
Timothy Jackson: No, and I—and again, I think it’s a super important point, just in the same way, on a much smaller scale, in a sense, but I mean with correspondingly massive influence over a shorter time scale, it’s important to me how we use the name Darwin, or whatever, you know, like. It is important how these names, when they are turned into their kind of patronyms—in this case, you know, they become misappropriations. They become distortions by, you know, through a kind of process of cherry-picking, or just simplification, or coarse-graining, or whatever of the richness of the founding text. It’s not that there’s no relationship. It’s that the relationship is so one-sided or so selective in an evolutionary sense. But, you know, it’s amplifying some very specific subset of the variation that was present in those original texts that becomes a complete misrepresentation because you don’t have all the caveats against making exactly that kind of move. And so I think that you should be even more concerned than I am about this so-called Platonic representation hypothesis in machine learning. And again, the kind of vulgar invocation of Plato’s name in service of something which is—well, yeah, I mean it’s—it’s I don’t know. I’m trying not to be rude, you know, like it’s pretty much…
Matt Segall: Bad science and bad philosophy.
Timothy Jackson: And I think things come together in the concern with, you know, rehabilitating the breadth of Darwin’s thought, and rehabilitating the breadth of Plato’s thought in an issue like this Platonic representation hypothesis, because, you know, the machine learning models and algorithms that are used to generate them can really be understood very lucidly in evolutionary terms. You know, they are a constructive evolutionary process themselves. They instantiate that kind of thing, and so it therefore requires a certain sort of literacy with evolutionary principles, in order to understand what’s going on, and then not leap to some limit which is the vulgar Platonism of, “Oh, what the models are discovering and pointing towards is this timeless universal space of semantics or space of geometries, or whatever.” You know, you’ve got to think in terms of things like shared descent homology, you know, convergent evolution like homoplasy, convergence in form and function and things like that, and have a deep time conception of how those things work. And maybe we can get into that in a little bit more detail, and we’ll maybe talk about this Platonic representation hypothesis and what it is in more detail. But we keep coming back to Mcquillan and keep sort of teasing these sorts of things. But yeah, no, I think your point is really important. I’m, you know, I’m convinced, and I don’t think I was ever—I’ve always been against the vulgar Platonism, right.
Matt Segall: Yeah, likewise.
Timothy Jackson: And given that that’s what—when people say Platonism like in mathematics, for example, like a lot of the modern usage of the term Platonism or Platonic space is coming out of kind of quote-unquote philosophy of mathematics, modern philosophy of mathematics. And I think that that is—yes, it’s recognizably Platonic, like. It has its recognizable roots in ideas that are in the Platonic corpus. But it is this kind of vulgarization of it.
Matt Segall: Yeah. Yeah. And you know, maybe we can dig into what Mcquillan says about the history of science and astronomy here—the changing ways that math was applied to an understanding of, to elicit some understanding of nature that, you know, really distinguishes modern science from Aristotelian science physics.
Where you know, Copernicus is depicted as offering this new Pythagorean-inspired Neoplatonic-inspired conception of the solar system with the sun at the center that’s rooted in appreciation for mathematical symmetry and simplicity, right? And so Copernicus’s model is geometrically simpler than Ptolemy’s, but actually not as accurate in terms of making predictions as Ptolemy’s was. That didn’t—you know, this heliocentric model wasn’t as accurate until Kepler added the elliptical orbits. And Newton came and added his understanding of gravitational—the law of gravity. And so there’s clearly something more on the level of an organizing idea or paradigm, right, in Copernicus rather than in the usual sense of science as an empirically driven enterprise. But what’s interesting about what Mcquillan’s doing here is that it’s in this example, the comparison of Copernicus and Ptolemy—it’s actually Ptolemy who’s more like the machine learning people in that he had a predictive model. And Ptolemy kind of black-boxed what was really going on.
Timothy Jackson: Yeah.
Matt Segall: You know, it’s like just trying to save the appearances here. How God makes the planets move and the stars, I don’t know. I’m just—this is an instrumental model that’s very useful because it does allow us to make predictions. That’s what the machine learning theorists or engineers are saying. Copernicus seems to be doing something else. And so it’s an interesting, you know, comparison for Mcquillan to make and to not note that because he’s thinking of Copernicus as more like the machinic Neoplatonists, or inaugurating this approach. But it seems that was Ptolemy’s approach, more so.
Timothy Jackson: Well, I mean the way you’ve characterized it is really nice, but I guess again they’re both taking different aspects of Platonism, because what Mcquillan is saying is that the Copernican modus operandi is to believe in the model, first and foremost, essentially, right, and then—and then to not force fit, obviously, but kind of make a fit with that. But I suppose, as you’re saying, you know, that’s what the Ptolemaic thing is as well. I mean, it’s what the Newtonian thing is, too, you know, and there are a lot of assumptions that are there in the background. I mean this—actually, I don’t know if we’re prepared to go there. But we did talk about the Perpetua Mobile a little bit last week, which was, you know, the perpetual motion. I don’t want to say perpetual motion machine, because this is where the issue becomes—the issue emerges the difference between a sort of perpetual, a process of perpetual motion and a perpetual motion machine, of which there are two kinds, both of which are outlawed by the laws of thermodynamics, right? But this is not really what, say, Novalis, as we were discussing last week, means by the perpetual motion. He doesn’t mean a machine. It’s precisely an anti-mechanistic conception. But there’s, I guess, a connection here in the sense that you know, Newton has certain assumptions which are carried over, and they’re broadly sort of Aristotelian assumptions as well, even though he’s departing from Aristotle in a lot of ways. He’s assuming a prime mover, for example, right? And he’s assuming a static background, actually, like he’s assuming a kind of absolute space and time. And then all motion is going on in it. And that leads to this obviously very fruitful, but kind of thermodynamic conception of—I mean, it’s not just a thermodynamic conception, but it’s there in thermodynamics—of a closed system, you know, with certain set of fixed boundary conditions and motion that’s going on inside it. But it’s really just a reflection of the boundary conditions. And of course, the eternal laws, and in that world perpetual motion becomes impossible, like the creation of a machine which again would have specific boundary conditions, and, you know, is constrained by certain eternal laws, you know, and the first kind of perpetual motion is outlawed because it is a creation of energy from nothing. So it violates, you know, the conservation law, the first law of thermodynamics, and the second kind is also outlawed because it’s, you know, a negentropic machine, an anti-entropy machine that, you know, never runs down because it can, you know, in an unlimited way, convert entropy into work. So it violates the second law. But this whole conception is based on that, you know, closed system idea, which is the—and then that, again, is the conception of the whole cosmos as a kind of closed system, and that is precisely one of these philosophical, metaphysical, obviously theological in this context assumptions which gets built into the method. And then the data essentially is made to fit that set of metaphysical assumptions. And this is how we get the mechanical philosophy—like it’s because we get really, you know, you and I’ve said this many times, but because we get really fixated on it being this reductionist whatever and, you know, anti-metaphysical in some sense way of doing things, but that claim of neutrality which is also attached to the machine learning things—we just forget, or it’s easy to forget, and it seems to often be forgotten in the broader discourse that it’s a fully theological worldview. You know, it’s a fully metaphysical picture, and it is enabled in a very important way by a set of metaphysical assumptions. And I think that’s kind of what—there was a bit of a digression there on perpetual motion and stuff. But I think that’s kind of what Mcquillan is getting at here, really, you know, not in exactly these terms, but that there are—there is this set of assumptions about the nature of, about the ontology. And maybe in the way you’re characterizing Ptolemy, it’s more like the way—actually, yes, Mcquillan is characterizing the data scientists here as saying, “Oh, it’s just a model, you know, like, we’re pure instrumentalists. We don’t have an ontology.” But again, that has typically been a way of covering over the tacit metaphysical assumptions. And what we’re seeing now, as we’ve discussed, is a more explicit embracing by data scientists of these, you know, allegedly philosophical implications of the thesis, but they’re not implications of the thesis at all, to Mcquillan’s point. It’s just people recovering their priors, you know. They were actually built in to the history of this way of doing things.
These vulgar, Platonic, metaphysical assumptions. And so now people are recovering them, and kind of going like, “Wow! Look what we found,” you know, like, and it’s like, “Yeah, that was just in there all along.” You know, it’s kind of like having a computational model of the cosmos and running on a computer and finding that it tells you that discreteness is fundamental. And you’re running it on a computer. You know, of course. Yeah.
Matt Segall: Yeah. Well, maybe we can jump from there, and it’s not too much of a jump to what Mcquillan draws out of Karen Barad’s agential realism. And how that, for me—point of view, I’m curious how it might differ from a kind of constructive realism, because, you know, Barad draws on Bohr, and I would—I had always understood Niels Bohr to be taking more of a stance of what I would characterize as like epistemological quietism where he would say, “Look, physics doesn’t tell us what nature is, physics tells us, you know, we get answers to whatever questions we ask.”
But all we really know, you know, it’s like we’re pre-stocking the pond with fish before we cast the line, and we can’t actually know what nature is independent of our line of questioning, and I took Bohr to be making an epistemological point as a kind of quantum Kantian or something.
Timothy Jackson: Okay.
Matt Segall: Whereas Barad reads him ontologically, and that’s a nice extension of Bohr, I guess, to make the realist revision she wants to make, but I don’t know that that’s Bohr’s position. And so I’ve always worried that drawing on Bohr for one’s constructivist reading of science leaves you with a kind of—in a Kantian situation. But I wonder how you read that? I don’t know if you’ve read Meeting the Universe Halfway, but…
Timothy Jackson: Bits of it. I’ve dipped into it. I got it a long time ago, as with, you know, too many books, and I was like, “This is clearly, you know, in my ballpark,” and I’ve dipped into it, and I may have watched a couple of talks of hers over the years. And obviously, you know, there’s what she has to say here. I wouldn’t say that I know her well. I mean I think it was Catherine Keller—in one of those process—we just watched about ten minutes or fifteen minutes of a process studies panel last week, or something from, I think that Whitehead…
Matt Segall: On the Harvard one.
Timothy Jackson: Conference, and it was either Catherine Keller or one of the other panel members, saying that I think it was Keller that, you know, Barad kind of disavows how Whiteheadian her position is.
Matt Segall: Yeah, yeah, that was Keller. Yeah.
Timothy Jackson: Yeah, yeah, which I thought was funny. And of course there’s a lot of Deleuze influence. And Keller was making the point that, you know, these people like Barad and others, are kind of getting their Whitehead through Deleuze. And so it—you know, it was interesting, because, the point being made there was that, you know, we can make—Whiteheadians can make interventions into the kind of Deleuzian literature, and part of that would be showing how influential Whitehead is on Deleuze. I’m like, “Don’t neglect Simondon,” because…
Matt Segall: No, that’s become very clear to me.
Timothy Jackson: Yeah, yeah, yeah, yeah.
Matt Segall: Many of the things I thought Deleuze was—that I got from Deleuze, are from Simondon, and like word for word.
Timothy Jackson: Yeah. Yeah. And I mean, I guess it’s the same with Bergson. And like, you know, Deleuze is a great—I don’t want to say, you know, synthetic thinker in, because he’s anti-Hegelian, but I mean he is. He takes a lot from a lot of different people, and he, you know, and but he very much does his own thing like. I think it would be silly to…
Matt Segall: He’s—but yeah, he’s doing a tremendous work of cultural appropriation from his own culture, which is great.
Timothy Jackson: I mean, he’s buggering all of these people. And he’s bringing in a lot of his own stuff, and obviously not to—to also allow the huge creative role that Guattari has in, in, you know, affecting Deleuze’s thought from like 1969. So. But yeah, Bohr.
I mean, people always joke that like no one understands what Bohr really meant or whatever. But I do—I think that you’ve characterized it, you know, in a way that’s consistent with my understanding, that I do think he was a kind of Neo-Kantian, and I think explicitly so. You know I think he characterized himself as such, and I think he is in that—and this, this is actually the kind of second-order cybernetics type position here, like von Foerster would say, you know, the environment is what it is. It contains no information, and we construct realities. We build our own realities, but it has no relation to, you know, no informatic or whatever meaningful semantic relationship to the environment as such, which is just neutral, which is really—I mean, it’s very disappointing to have that kind of view from within a—because it’s also there’s a performative contradiction, because the, you know, in acting as instrumentalists, these people do tacitly assert a realist aspect to their philosophy, but then they disavow it, which is kind of like disavowing the consequences of your actions. It’s like, “Oh, well, the environment’s just there, you know, I’m not really changing that. I’m only interacting in this very superficial layer. So it doesn’t really matter in some sense,” and there’s a lot of then appeals to—again, it’s vulgar, because you and I would both want to defend a certain notion of mysticism and esotericism, and all of that, and I mean again, it would be broadly Platonic in a lot of ways, but not exclusively. There’s a kind of vulgar mysticism associated with these positions as well, and not—not that I want to—I mean, maybe I do want to hack on Carlos Castaneda a little bit, I mean. First references Castaneda, a bunch—actually Deleuze and Guattari reference Castaneda a bunch. And there’s a lot of, you know, validity in things that Castaneda says, but if you look at the way Castaneda’s own trajectory plays out, I mean, not only did he make up Don Juan—so far as we know there is no, you know, Yaqui Indian, that he was inducted into this, you know, Shamanic tradition by, he writes a bunch of clever stories which are obviously very well researched and full of insight. But then he becomes a cult leader, and you know, people die on his watch, as, you know, literally. So, anyway? So there’s there is a potential danger, though, in this kind of anti-realist disavowal. Okay, we were talking about Karen Barad—and not attributing that to her. Yeah, I think, she ontologizes. It seems to me that she ontologizes this. She goes from Bohr, but I think that that’s kind of consistent with sort of relational quantum mechanics. And you know, people are working out of a kind of—it doesn’t need to be Bohr per se. But it’s going to be very structurally related to Bohr’s position, because he is so influential in, you know, shaping everyone’s early understanding of quantum mechanics.
And so, you know, it’s not surprising if there’s a lineage of dissent there, and that again, you can always turn this kind of Kantian—I mean, it’s like Schelling, or whatever, you know, you can always turn this kind of Kantian position, or indeed, this Platonic position can go in the two-worlds way, where there’s a real schism, or you know, you can occupy that middle ground and build bridges. And I think what you know Barad is doing is occupying that middle ground and building bridges, you know. So I, you know, agential realism—I’m like, “Yeah, of course, that works for me.”
I mean, there are things that are said here, like in Barad’s account, phenomena are these inseparable physical-conceptual interactions. But it’s like, that’s just what phenomena are, you know, like it’s not—that’s Kant’s account as well, right? Because there is the noumenal and then there’s the conceptual, and, you know, these things are interacting. And that’s what we are perceiving, essentially.
Matt Segall: I mean to the extent that what I see these process-relational insights coming through Barad’s articulation of this, I’m on board, but it’s just that connection to Bohr, where I would describe Bohr as an agential anti-realist.
Timothy Jackson: It’s easily invertible, isn’t it? Like these things are…
Matt Segall: Apparently so. Yeah. Because I mean, both of us want to break from both the von Foerster constructivism and our constructive realism. We want to break from classical, representational science and that kind of naive realism, right? And so to the extent that we’re resisting naive realism, the radical constructivist position and the realist constructivist position, I think, seem similar, but are actually—wanting to ontologize something that otherwise just remains an epistemological sort of skeptical move.
Timothy Jackson: Yeah. Yeah. And I think again, I think there is a place for this epistemological, skeptical move. And we do have mediated access in a sense, right? But again, you’ve got to put that back into a relational realism, basically. But to your, I think to your point. And one of the things I’ve got underlined here on, and a little note on page—it’s page six. So it’s page six of the paper—he’s got a quote which I assume is from Barad. It’s actually not—I mean, he mentions Butler earlier in the paragraph, but he says, “Agential realism describes participation rather than reinforcing dualism”—quote not cited this quote, but I assume it’s from Barad.
“If our descriptive characterizations do not refer to properties of abstract objects or observation-independent beings, but rather through their material instantiation in particular practices, contribute to the production of agential reality. Then what is being described by our theories is not nature itself, but our participation with nature. But the whole point is here that there is no nature itself outside of participation.” So it’s like, again, I’m not—I’m only picking on a particular way of phrasing here. I’m not saying that this is her view, but that is, in a sense, a reinstatement of the kind of Bohrian, or, you know, naive Kantian, or whatever position where there is a schism here—there is nature in itself, and then there’s this participatory layer.
We exist, and—you know, again, you need that Schellingian inversion which you call “descendental,” which is like the whole thing has to have a genetic account, you know, in order for that participatory layer to exist, and we can still acknowledge the reality of a certain kind of mediation, and like a perspectivism. But in order for that participatory layer to exist, you cannot begin with like a pre-constituted subject and a pre—I mean. And again, this is, he says, is the sort of the point of Barad, but you’ve got to take it all the way back.
You know, the subject has to have a genesis which is embedded in the world, and it’s its genetic account. And he says, “Agential realism does not presume specifics about the world prior to the enactment of material-discursive practices.” Well, I think that that’s got to be wrong. That it has to, because if it’s going to be a thorough-going ontology, it has to. It can’t be that material hyphen discursive practices are like absolutely prior. They have to be themselves derived from some, you know, again, onto-genetic process. And so you are—like, this is the discussion that you and I are always having about, you know, it is necessary to think just as a limit a kind of primordium. You know. You have to think of like what would it take? And how far can I reason back to some minimal model? What would it take for the world to have these kinds of properties for material-discursive practices to emerge in this, you know, fully participatory way, and whatever, you know, something that looks like that has to kind of go all the way back and not to seek a determinate origin story, but to recognize what are the operators that you’re going to need to have an ontogenetic or developmental or evolutionary account of, in order to achieve this, you know, sophisticated, agential, realistic, you know, interaction of material-discursive practices. And I just—that’s the neglected thing, which is, again, why these positions kind of appear to come apart. The anti-realist and vulgar realist positions when they’re, in fact, kind of instantiations of the same position, because what they’re both saying is like there is a reality, and it’s kind of absolute and permanent and static, and we—but we don’t have access to it, you know. So the machines have some kind of access to it, or the Demiurge has access to it, or whatever—like it’s, you know, when pre-fall we had access to it. Pre-incarnation, or whatever, and I mean. Von Foerster is saying the same thing, you know. That vulgar second-order cybernetics position is, and anti-realist position is saying the same thing.
Environment in itself has no information, has no meaning, so it has—genetic, it has no genesis in that case, right? And then—and then I think also, the subject has no genesis. So I think, like this is what subjects do. The environment is there and subjects do this thing where they create worlds, so that the world—the act of worlding has this beautiful genetic account, and it’s like a sophisticated theory of abstraction. And it’s great. It’s really good. You know, you can tap someone like von Foerster for that. That’s really, really powerful and needed. But you—it assumes a subject that does that, and a world that does nothing. And then and then it’s like we go from this point. But it’s like, you know, you have to generalize that sophisticated account of abstraction across both poles, and then you have—well, you have a process-relational ontology. It just—it’s just like harder to do that because you’re taking way less for granted.
Matt Segall: Yeah, yeah, I mean, this is all reminding me of Schelling’s reading of the Timaeus as a transcendental exercise, and you know the way in which—I love that you point this out, that you know it can’t possibly be the case that agential realism doesn’t presuppose about the world prior to the enactment of material-discursive practices. I mean the whole account in the Timaeus, especially as Schelling is reading it, is like, how did we ever come to a situation where we need a hyphen between material and discursive like? What is this relationship between the sensible and the intelligible, that we can’t escape in our practice of science now? Like, how did that originate? And Plato goes through this very anthropocentric idea, rooted in a material practice of, like, you could say, pottery-making. You know, the demiurge shaping the chora by reference to these forms. It’s so obviously anthro-centric, and it’s not to be literalized, you know, and the way that the Gnostics end up literalizing it. But it’s an exercise in trying to think ourselves—to think the generative process, the process of genesis by which we came to be in need of a hyphen between the material and the discursive, you know. And so like. There’s a hidden metaphysics behind machine learning. Yes, and agential realism is a metaphysics, and so is making assumptions—speculative posits about the world. You know.
Timothy Jackson: And it presents itself as a metaphysics, and he’s presenting it as a metaphysics, and he’s presenting it as an alternate metaphysics. And I think that’s all really good. But I think again, it just isn’t going far enough, at least in its presentation here. Human practices are—hang on. Yeah, human practices are agentive participants, as phenomena—as phenomena are sedimented. Sorry. This is terrible. Are sedimented out of this ongoing process. “Agential realism is the notion of material-discursive practices that produce the world through a process of sedimentation—that is, the iterative layering of phenomena that produces subject and object rather than taking them as pre-existing entities.” So it’s like, some of that is exactly bang on. And some of it is like, you know, not going far enough, because, yeah, it’s these processes of sedimentation or stratification, as you know, Deleuze and Guattari would say. But I mean, as geology would say, right? So you know, “Geology of Morals”—great chapter in A Thousand Plateaus that I think we should read, because I think it was really great evolutionary thinking. We should read it together. But “you know agential realism is the notion of material discursive practices that produce the world”—well, what about the processes that produce the material discursive—exactly to your point.
Matt Segall: Which, by the way, all that is such a great description of what Iamblichus is doing when he—divine names, and using stones and plants and other sacred figures, you know, to evoke the divine. That’s precisely what he’s engaged in.
Timothy Jackson: Isn’t that just what ritual is in general, right? And I think, you know, I’m not a big fan, necessarily, of Aleister Crowley, and I don’t claim to have great knowledge of him, but you know I read Moon Child many, many years ago, and his major novel, and I really enjoyed it. It’s a good novel. And he has this point, which I was like, “bang on.” You know, he’s talking about someone engaging in—I can’t even remember the details, but like an invocation of the spirits, right, and some sort of magical ritual. And he basically says, and she or he, I think it’s a she, but it doesn’t matter—never stopped for once—never stopped once to consider whether this was all real or like, you know. But it was efficacious, you know, like it was having an effect in the world, and that was what mattered. And but again, it’s really interesting, isn’t it? Because this is how instrumentalism and mysticism are so closely related, and how magic is, you know, in the—I hate saying in the DNA—like, but you know, in the inheritance of science, right? It’s in the genealogy of science. And so it’s weird that we get this like—the science as this kind of breakaway reactionary movement which is trying to escape from dogma and all of that. And of course we know, you know, it’s very, very dogmatic in many of its own terms—can be, although you know, I would always say science is a practice, not a body of knowledge, you know. But scientists often are very, very dogmatic. And one of the things that science as a worldview is supposedly deeply against is magic, like, you know, that’s supposedly what naturalism is. Naturalism is supposedly the “no miracles” argument, right? And I mean, I’m good. I’m good with no miracles in a certain sense, right? But it’s like magic is not about miracles, you know. It’s about…
Matt Segall: No, no, it’s not.
Timothy Jackson: Entering into—we talked about it last time—this kind of sympathetic resonance we…
Matt Segall: Humans do magic. God does miracles. It’s very… But yeah, I mean, I think there was a transition in the 17th century, and it’s very explicit in like Francis Bacon, between this older form of sympathetic magic that was associated with witches. And you know, herbalists and people closer to this occult understanding of signatures in nature, and a transition to a different kind of magic which wasn’t sympathetic magic, but maybe machinic or mechanical in the sense that it’s seeking not to be in sympathy with the natural—say the Logos of the cosmos, but to use human-made mathematical models to sort of gain power over nature, right? And so it’s—I don’t know what the right term for the opposite of sympathy here would be. But there’s yeah. That shift from sympathetic magic to a kind of instrumental magic, perhaps.
Timothy Jackson: Yeah, yeah. And I guess you know, I want to highlight on the bottom of this page of six—what I think is kind of a misunderstanding of techne. But in a way, what he’s characterizing here as techne is that instrumental—is that instrumentalism of like a kind of power and dominance, and again, to go back to the perpetua mobile, in a sense—like this, this is the division of different conceptions of perpetual motion as well—that there’s certain kinds that you like. You can’t make a perpetual motion machine in these ways—you can’t master, you know, energetic flows in these ways, but that’s very different from saying that the cosmos itself is not a perpetual motion machine. But what happens is that with the mechanical philosophy there’s a conflation between the machine, the technical object on the table which can’t do certain things, and the cosmos as a whole, you know, or what I would call like steam engine metaphysics. And that is precisely this distinction, again, between the different conceptions of perpetual motion in like the romantic tradition and the mechanical.
And this is very related to the refutation of magic, because perpetual motion machine becomes a kind of quintessential example of magic of the bad magic, you know, in the mechanical philosophy—like this is magical woo-woo stuff, you know, but it’s like that’s really not what the poetic image of the perpetual motion, and this again entering into the process of like reciprocal becoming. And the reason I mentioned it earlier is because when he talks about the shift to, you know the Ptolemy to Copernicus, and then gets to Newton. And you know this Newton—this inertial motion? Which kind of sounds almost like a kind of perpetual motion, in a sense, right, like the natural state for things is to be in motion, in fact, until they are constrained, or that motion is interrupted, and that’s sometimes presented as kind of the antithesis of Aristotle and Aristotle’s belief that the natural state of things is to be at rest, and so they need to be pushed, moved, and hence blah blah blah prime mover. But, Newton—I already said, you know, Newton fully endorses the prime mover, and he has the static background, and all motion is relative to a static background in Newton. And what Mcquillan says is that this is a replacement of the older vision, where motion was part of the becoming of the thing itself, right, part of the becoming of the cosmos. And again, that’s precisely this distinction. Obviously, you and I want to want to rehabilitate and recover a notion where motion is itself a manifestation of this becoming, which is like a kind of motion. I mean, you’ve told me off in the past for saying that, and we don’t need to get into the differences between like intensive and extensive motion and stuff right now. But it’s a kind of moving—an animacy which is fundamental and kind of the prosaic motion that we see in the world all around us is a manifestation of that deeper, you know, generative principle of just like not being at rest, you know, like rest is an achievement—not necessarily a good one, but it’s a function of constraint. And then we can get into, you know, metastability and energetic landscapes, and all of those really useful thermodynamic, because there’s tons of like gold in thermodynamics and statistical mechanics. Obviously, again, without needing to reify certain things or elevate things to like, you know, universal rules which are clearly emergent and emergent appearances, in fact. So there’s that distinction between appearances and reality. But at the end of this page here.
He says, you know, “practice is more than a praxis.” You know. So what is he saying? “Data science”—oh, he’s saying a counterculture of data science can be a critique that also becomes its own practice. So, like the machinic Neoplatonism of data science, which is a praxis which is an acting directly in the world is an exertion of force, power, capacity, etcetera, and a direct changing of reality, not just a representation of it, but.
Matt Segall: Exertion of corporate and state and military power. Yeah.
Timothy Jackson: Well, exactly, and hiding behind this representational notion. This is exactly that cusp of this, like really dark alliance between anti-realism and realism. In these, both in these vulgar senses, that on the one hand, it’s like, “Oh, it’s just representational. Don’t worry about it,” you know. But, on the other hand, it’s like hyper-real and machinic, not just mechanical, acting directly in the world. But he basically says that, you know, the critique can become a counter-practice, right, so critique also becomes its own practice. In other words, a machinic form of praxis—machinic form of praxis is praxis always machinic, I think. Probably. Yes, but he says, “practice is more than simply reflective practice.”
Again, reflective practice is changing things in the world. But I get the point because it also contains the idea of the good that is an overall goal of human flourishing. Okay, very Platonic, right? As I know. I see your face lighting up. You want to say.
Matt Segall: Well, he didn’t capitalize good, but well.
Timothy Jackson: Know. But he says then, “Instead of techne, a way of being concerned with making things, and with what things can make, practice is a political action as a mode of togetherness.”
I guess that’s interesting. I mean, maybe it’s not fully a mischaracterization. But I think you know, techne—techne has a lot of different aspects to it, and I think one of its key, you know, one of Aristotle’s key definitional points about what the nature of techne is is that it is a form of mediation. So it’s a mediation between man—but I will extend that to like organism and world. And I think of, as you know, and anyone who’s listened to a lot of our conversations knows, I think of niche construction as the kind of fundamental form of techne. In fact, we talked about this last week, and this romantic mechanology, and all of that, and that, I think, was the view, without saying it in exactly those terms that was being articulated in that paper on Simondon and Novalis. You know, this kind of situated participatory mechanology is a vision of techne, as, you know, participatory, but part of just how organisms shape environments by being in relationship with them and kind of unavoidably so. You know. There is no view from nowhere. Blah blah blah blah. So I think maybe he isn’t quite mischaracterizing techne. But there’s an opportunity there to kind of say, well, techne is always already this kind of acting in the world and shaping the world, you know, in our own image, I mean. I don’t know if that’s if that’s where I want to go with it. But I mean in niche construction of any kind is technical and sort of life is inherently technical, because life, in order to persist, has to control at least some of the degrees of freedom of its environment. That’s technical. That’s technicity.
Matt Segall: Yeah, yeah, it’s making me think that maybe you know, Heidegger’s distinction between techne and poesis as different modes of revealing. And he says, technique challenges forth and poesis, let’s be. And let’s what is reveal. Or you know. That is similar to the distinction I was trying to make between the sympathetic magic of the pre-scientific age. And then the form of, you know, the way magic transmutes into this new instrumentalist mechanistic science, where there is more of that technological interventionist way of engaging with the nature where it’s more important that we force fit the data to the model that we prefer because of its symmetry and simplicity.
That this is a way in which technique goes wrong. But at the same time you could see poetics and organics as a form of techne as well.
Timothy Jackson: And yeah.
Matt Segall: And it’s always a collective endeavor. You know, the technology is this whole point is technology is not value neutral. It’s embedded in these cultural narratives and practices that give it meaning.
And he and he taps Foucault for that point here. Yeah. Social apparatus. Yeah.
Timothy Jackson: But it’s a Simondonian point about technical objects, and probably it’s partly from Canguilhem, and, like, you know, Foucault is definitely very influenced by Canguilhem, and so is Simondon. But I think you know I wouldn’t be surprised that Simondon was also influential on Foucault to some degree, because I think Du mode d’existence des objets techniques was kind of a big book for a little while in the French philosophical circles, and that was published in the fifties. So yeah, so yeah, you have something to say. There was a couple of things that I was gonna say, but I’ve lost them all of a sudden.
Matt Segall: Well, I mean just this, the challenge to what he calls onlooker consciousness. With an overt mission to manipulate, as he puts it, and that you know, the connection is that this is the—this is a derivative of Neoplatonism in some sense. And you know, obviously the whole idea of the demiurge beholding the forms and shaping materiality on the model of the forms—like I get. Why, you would say, “Oh, that’s where the onlooker consciousness comes from.”
But at the same time, you know the methexis—this idea of participation, I think, is precisely about—it is, you know, in Platonism is also the more participatory ontology that he’s calling for at the same time. And so again, it’s like we’re painting with this Platonic palette. The positions that we nowadays grapple with that seem opposed to each other are both building on what the interlocutors in these Platonic dialogues were arguing, you know, the positions that they were arguing against each other about and so I think that we can look to the same corpus for the alternative to the onlooker consciousness, which is this participatory ontology that he’s describing?
And you know, in a participatory ontology, science is just as much an art or about technique as it is about discovery, right? It’s just as much a creative endeavor where you know it’s not that the world is nothing until we make it meaningful through our practices and applications of forms of measurement, and whatever—there is a world out there that resists us and our interventions into it.
But the goal of science is rather than manipulation is to come into communion and to enter into harmony with the world has not—yeah. A neutral.
Timothy Jackson: Hmm.
Matt Segall: Just neutral noumena, as it were, waiting for us to project onto it. You know.
Timothy Jackson: Yeah, yeah. I guess, on the onlooker consciousness and a kind of Platonic, but not—not exclusively. I mean, it’s very much there in like stoicism. But they’re in Buddhism and like there’s a—there’s a soteriological practice aspect as well of the onlooker consciousness, which is the sort of what Pierre Hadot would call the view from above, right.
Matt Segall: But it’s the look, but don’t touch. That’s the soteological is the look, but don’t touch approach not the look and manipulate.
Timothy Jackson: No, for sure. Yeah, yeah. But you know, transcend the fray, right? So yes, exactly as you say, you know. Observe, but don’t react, you know, don’t interact. And I mean, that’s a really important practice for sure. And you get to see a lot more because it’s also—it’s an emptying, right? You get to—because you’re not, you know, engaging through those instrumental modes like, “How can I manipulate this thing?” You’re just observing it? You get to see a lot more. And this leads to, you know, a notion that I call sacred nominalism, and all of that, like a meditation on firstness.
Which has its own relationship with first—first as idea that the environment is just what it is, you know, like, there are, there are all these structural relationships. It’s just that they are not—and there are. There are virtues to those views, but they’re not self-sufficient, right? And they’re not ontological, you know, like a meditation on firstness doesn’t reveal to you that “oh, the world is just what it is, because I can step back as far as possible and have as minimally mediated”—it’s still pretty damn mediated, because you still like the kind of organism that you are perceiving things through your sense organs and all of that.
But just because you can step back out of a particular set of like priors, or, you know, layers of mediation, some of which may be instrumental, doesn’t mean you’re getting to like the ultimate nature of reality. And I mean, this has been a big issue in, you know, mystical, contemplative traditions for a very, very long time. And so there’s also, on the one hand, there’s this—what is it like? Reification of the noetic state, you know, like as a naive realism in translating your experience of noesis into an ontology or an interpretation of the ultimate nature of reality. And that’s why I have, you know, the prosaic, the, you know, inquiry, and speculative as kind of different stances, in a sense, that you can take to different modes of interaction with reality.
And as soon as you say anything. As soon as you enter into the discursive about the nature of some experience that you had, you’re already out of that, you know, like legitimacy that the experience has of itself in and of itself. But there’s a notion that there’s like a pernicious notion of transcendence here as well, right. So again, if you think it is some kind of absolute transcendence going on as opposed to like transcendence within immanence—like you’re still really embedded. You are just, you know, when really you’re still acting in the world.
Matt Segall: I’ve often—I’ve encountered. Well, the nonviolent communication crew is also very argumentative, but the non-dualist crew is super argumentative. They find reasons to distinguish their view from other views all the time, which is ironic. But just to say about firstness and your sacred nominalism. It’s kind of sacred—you can’t say anything bad about it. No. But it’s ironic that you’re invoking a triadic ontology. Firstness implies secondness and thirdness, which for Peirce was an example of his realism. And so I think, while firstness might itself point to the sense in which no name is adequate, I think you know a realist would agree. But if there’s firstness and secondness and thirdness, there seems to be something more going on.
Timothy Jackson: That’s the point. That’s the point. You know it.
Matt Segall: That’s the sacred part.
Timothy Jackson: Well, but that’s—I mean, so you’re taking exception, basically to the use of the term nominalism. But what I’m doing is I’m saying like, “well, that has its place, and I’m sort of bracketing it.” I mean, I don’t like -isms, anyway, because they just seem to absolute. They’re like absolutes in a sense, right? So we don’t have to call it sacred nominalism. Whatever. All I’m just saying is that there is a—there’s a place for that nominalist intuition. And it’s like—here in my scheme, anyway, it’s like, I also trap God in the prosaic, you know. I’m like, “Of course there are, you know experiences of God, and of course, that they are, you know, absolutely inviolable, and all of that.” But they’re quite different from a spectrum of ontology which has a role for a God entity. But yeah, I mean I—it’s totally fair for you to take exception to nominalism there, but it’s whatever—I mean. It’s just a catchy little name. I don’t really care.
Matt Segall: Holy Trinity better, but you could say sacred, not awesome.
Timothy Jackson: That is a conception of a whole of the triadic scheme, whereas, like, you know, sacred nominalism is more like a meditation on the Holy Spirit, because it’s a—it’s a kind of meditation on that creative principle, which is, you know, first first, because it’s generative, because it’s spontaneous, in fact, because it bloweth where it listeth, or whatever, you know.
Matt Segall: I’m very happy with that.
Timothy Jackson: Yeah. Yeah. So that’s kind of what I—that would be my equivalent. Yeah. And so that’s where these are stances. And the stance towards firstness. And it’s not just a stance that, like meditating humans can take. It’s a stance that, you know, nature immanently takes on itself. For example, when one level of an organism—you know this is a vulgar way of putting it, but treats, you know, the spontaneous gene expression at a lower level—like a molar level, treats a molecular level as spontaneous and as continuous. So there are continuous distributions of molecules. In fact, the molecules are discrete. They’re actual, but—and there are boundary conditions and all of those things. But there’s a view on them where there are continuous distributions, and those continuous distributions are subject to the possibility of canalization of differential canalizations. And so, whilst there are actual molecules in play, this is the virtual aspect of spontaneous gene expression. And so it’s taking a stance towards it in its, you know, firstness, right? It has a—it has aspects of secondness, you know, like molecules are like banging around. And of course it has aspects of thirdness, because it’s an evolved, you know. It’s not like the primordium is not a bunch of, you know, genes—like the cosmic primordium is not a bunch of, you know, gene products bouncing around. But there’s something that—there’s a principled similarity between what you’d be wanting from a primordium and this kind of spontaneous, you know, uncanalized from a particular perspective, you know, aspect and all of reality is kind of like that. So all of the actual has its virtual, you know, component. And that’s like its exactability, you know, that’s like a technical object is like a hammer is not just for hitting nails, you know. It can be—well, it could be used as a weapon, or what—you know like. That’s just a nasty example that’s coming to mind. But you know things are multiply realizable. They are exaptable, you know. I could stick this pen in my ear and get relief from an itch, you know, which is another not very nice example. But you know what I mean right, and that would be treating the pen, not in its aspect of thirdness, because its thirdness is its purpose for writing, but it just happens to be an object in the world which has all these other kind of quasi-phenomenal or incidental properties? And I can…
Matt Segall: Exapted in a variety of ways.
Timothy Jackson: Exactly so. This is just a theory of exaptation. There was a connection I was going to make before, and now I’ve kind of lost it. But I’ll just jump to it, anyway. Because it’s just about this participatory aspect of everything as well. Back on page six, the. You know. The stuff that we characterize through—so he’s talking about agential realism. This is where he’s just begun to introduce it. “The stuff that we characterize through experiments cannot be said to exist in that defined form between and independent of us measuring it.”
This breaks with classical representational science. Bohr talks about phenomena as particular instances of wholeness, the inseparable object-measurement event. So in a really interesting sense, because this is a quantum metaphor. And so we’re saying that there’s like a kind of a smeared, relatively indeterminate, or like probabilistic distribution of possible states. And then in the act of measurement, then we define reciprocally, not in a deus ex machina way, but by setting up an experimental apparatus in a particular way, and then, you know, constraining a set of potential which are not infinite. So we’re not just creating something out of nothing. Then we get this determinate outcome or this defined form. But I guess the point to make here just is our same refrain is that, you know, the stuff that we characterize through experiments cannot be said to exist in that defined form between and independent of us measuring it.
If you’re generalizing that from the kind of quantum—it’s just important to remember that stuff is measuring everything else all the time. So like classical reality is like massively entangled. And that’s why there is structure, because you have constraint, you have coupled degrees of freedom which are mutually constraining. And that’s how you get stability. And that’s why things have like determinate and relatively stable over particular time scales, including like the length of the entire cosmos stability. That’s why you have structural stability. So it’s really important that we don’t leap from the quantum to consciousness here, which is again what always goes on. And it’s just—it’s so rampant in the current discussion. As well, you know. I mean, even the machine learning—conscious machine learning, even the machine learning discussion. It’s like, “here are these amazing, you know, algorithms doing these things. And are they conscious?” And it’s like, but you know, there’s this all this reality here, you know. And if you want to understand consciousness or machine learning, or whatever, maybe thinking about all of these other layers of reality, and how these constructive processes are going and how we get structure in general, and how humans don’t need to be in the loop because there was a hell of a lot of structure around before there were humans doing like bloody quantum mechanics—like this becomes a really important aspect of this. And again, I haven’t read Barad, so I don’t know to what extent she goes into like, you know, evolutionary theory and blah blah, but I haven’t read her in detail.
But again, this is the key point for me. It’s like, if you want to be in, you’ve got to be in the—it’s another place you’ve got to be in the middle ground. It’s not quantum mechanics and consciousness, it’s reality, you know, and all these ways, that it is self-measuring.
Matt Segall: Yeah. And that’s, I mean, this is a, I think, inheritance of Cartesian science, where it’s this…
Timothy Jackson: Yeah, yeah.
Matt Segall: Split between mind and matter. And you lose the mediating term which was present in Plato. Life. Life is the mediating—soul is the mediating term. Schelling revamps that—all the Romantics revamp that—you know, it’s Schelling’s always harping on about the lack of mediating concepts in mechanistic science, right?
Timothy Jackson: Yeah, yeah.
Matt Segall: And that’s, you know, this leap from quantum physics to consciousness is “what collapses the wave function,” and that there’s no interaction happening until a conscious human being, you know, checks up on the cat, or until you know, conscious attention is able to collapse the wave function is—it’s just a really—I don’t want to be ableist. I was gonna say, a handicapped mode of thought. But it’s like it’s really—we need a bridge. You know, we need to remember the bridge between matter and mind. These abstractions is the living world and the biological world and the evolutionary process. So that’s a really important point you’re making. And yeah, like, of course. Facts, the facts, the data that’s salient to us depends on the frames, you know, that we’re imposing on the world. But your point is that nature is self-framing in a way—nature coarse-grains without us having to do it.
Timothy Jackson: And again, it’s an account of the genesis—onto-genesis of these systems like conscious human observers, right? They know.
Matt Segall: Right? Yeah, yeah, of course, historicizing the transcendental is.
Timothy Jackson: Well, exactly. And I mean again, that’s the Schellingian.
Matt Segall: Schellingian move right as you as you have so eloquently, you know, pointed it out.
Matt Segall: Just got it from Plato, so.
Timothy Jackson: It’s a move to a degree, I mean, I do think there’s originality.
Matt Segall: No, it’s a creative reading, but I think, you know, there’s so many layers to the dialogues.
Timothy Jackson: No, I agree.
Matt Segall: Intentionally and probably unintentionally. You know. That’s the way the unconscious works. I don’t even think Plato knew everything he was communicating in these dialogues.
Timothy Jackson: But maybe Plato felt like Socrates, that the only thing that he knew is that he knew nothing, you know, and that’s why he wrote dialogues in some sense, right, as well as to defend himself from posterity. Because, you know, again, there’s the immanent critique of writing as Pharmikon, and you know I’m not going to be there to defend myself, so I better put—you know I better put all these theses in the mouths of interlocutors, and kind of try and present every point of view. No, it’s and it’s brilliant, obviously. And these ideas have happened. So there’s something very contemporary like—not chronic but timely about Schelling’s thought. You know, I mean, it’s untimely in the sort of Nietzschean sense in that we’re still, you know, processing this, and it’s taken a Darwin and others, you know, subsequent to Schelling.
And again, not to see Schelling just as this originary figure either, but to develop contemporary evolutionary thought, and these detours through kind of instrumentality, you know, and the kind of bastardization of neo-Darwinism, and all of that, like molecular biology has a huge amount to offer to this discussion. The molecular biologist says, but you know it was itself all framed in a very, you know, reductionist—like in a regressive way.
But you know these ideas have their times. And so Schelling’s thought is very much in the air in the sense that biology is a nascent science in the 18th century, and it has a huge influence on Kant. You know, and on the structure of the third critique, so that whole, you know, lineage of thought—critical philosophy is very timely, and is itself a response to many things, not just to biology, but in part to biology, to the nascent science of biology.
Matt Segall: Yeah,
Timothy Jackson: One of the other things I wanted to touch on on page—it’s seven. So it’s page seven of this, and it’s where he references Donna Haraway and the view from nowhere, I mean, did she coin that term? That’s a pretty good comment, like Nagel has that term.
Matt Segall: Yeah, I don’t think she—she didn’t coin it, but she uses it a lot as a whipping boy.
Timothy Jackson: So I mean he—he says there’s a few things here. There’s one thing that I that I kind of want to get to. But I mean, yeah, counterculture of data science refuses to throw the baby out of the bathwater. So it doesn’t abandon the idea that empirical and mathematical models of data science can generate valid propositions. Yes, obviously good. But data science is slippery on the point of objectivity because it’s not claiming to discern objective reality. It operates through forms of mathematical and computational objectivity. That’s where—this is not the case necessarily so much now, that’s where these Platonic representation hypotheses and related ideas seem to be much more strongly claiming that the machine learning—not that the models are capturing the objective Platonic patterns so much as when you map kind of apophatically the commonalities. So this vector—Vec model vector to vector model. You know the commonalities amongst embedded space representations of semantics.
You do a comparative—this is very evolutionary. You do a comparative study in evolutionary science. We would say, there that enables you to infer a common ancestor, right? But they’re saying that it enables you to infer the universal, timeless, Platonic, you know, ideal space. And it’s like again, a little bit of evolutionary literacy, and I want to get on to that before we finish. Would be nice here, because there’s a big difference between inferring a common ancestor, and inferring something eternal.
Structurally there’s a relationship there, but there’s a still a very big difference. But then the view from nowhere and then Haraway’s call for situated and embodied knowledges—form of operating that takes embodied responsibility. And this is the point that I kind of wanted to get to. Moreover, this embodiment should start at the edges. Standpoint theory proposes that positions of social and political disadvantage can actually become sites of analytical advantage because they can challenge hegemonic assumptions.
I just think this is—I’ve written “heterogenetic loci” here, and I just think this is a nice point. I mean, there’s a political point here. But again, there’s this same kind of evolutionary point about these differences. These loci of difference, which are not absolutely canalized by this, you know. What does he say? Hegemonic assumptions—they can provide? They are. They are grists for the mill of the evolution of novelty, essentially, right, precisely by virtue of the fact that they have a somewhat independent evolutionary history, or that they have a certain virtual, or, you know, firstness aspect to them that can be exapted, I mean, exaptation and co-option are actually terms that are used somewhat interchangeably, particularly in molecular biology. Politically, we’re going to want to make a distinction there because we don’t want it to be like a co-option in the sense that it’s the hegemonic assumption, just, you know, incorporates this view, and you know, neutralizes it, and it’s just, you know, business as usual, like, you know, the corporate co-option of psychedelics, or whatever we want it to be disruptive and—but co-option in molecular biology is the result of something. It results in a process which results in something genuinely novel, right? And that’s the thing that we want to highlight here, I think.
Matt Segall: Like that point as a general point—yeah, yeah, so that you know again, the participatory science, would recognize, I mean, I wonder what you think about this when we build something like a particle accelerator. We’re obviously, you know, we’re spending billions and billions of dollars to intervene into nature in some way, and to create conditions that we analogize to the origins of the universe, but that are heavily, technologically mediated. And it’s almost like we are. I think of it by analogy, to like trying to extract information from somebody by torturing them.
You know, you’re not necessarily going to get the truth out of them. In that way. You’re gonna get something, but it’s gonna have a lot to do with the torture method, you know, or you know, it’s not a perfect analogy, but it seems like with a particle accelerator. There’s so many theoretical assumptions built into the design of the apparatus that there are only so many results we could get when we turn it on.
And so it’s heavily constraining. What nature can, how nature can answer the question that we’re posing to it, and then to go and say that “oh, therefore nature is this way” seems to neglect the participatory dimension, and to reify this onlooker form of, or to get to be stuck in this onlooker form of science, despite the obvious technological means, you know, that one has to go to to get this data.
Timothy Jackson: Yeah, no, I think that the point is really well made, I think, because what I don’t hear you saying is that this isn’t, you know, a viable method of generating really deep insights about the nature of reality, because I think it is but you are saying that it’s mediated by a certain—again, we just hop on the same refrain—kind of history of ontogenesis, which is everything from, you know, these, you know, technical lineages of apparatuses. But the human—I mean our entire evolutionary history. Again, we’re trying to reconstruct a primordium from a very situated in evolutionary history, and we need to attend to all of the intervening layers of reality. Or you know, this temporal hierarchy of enabling constraints that enable us to get to this point, but it might be a point from which we cannot see everything. I mean, it is a point from which we cannot see everything, you know, and so I always talk about how we’re inferring ancestral states from extant variation, which is exactly what’s going on here.
Matt Segall: But this isn’t just a point about like the historical contingency of science, and that, like, you know, maybe intelligent animals on other planets would have a totally different physics just because of the way their history of knowledge unfolds, but also that it could be that we’re actually creating new particles that haven’t existed anywhere else in the history of the universe, and that this apparatus is generative. It’s not just discovering what’s there, you know.
Timothy Jackson: Yeah, no, I think that that’s a really great point. And I mean, we can analogize that to one another point that we often make, which is the creation of closed systems, the thermodynamic sense, and the way we are obsessed in the evolution—you know the human sociocultural evolution of the science of thermodynamics with constraining energy to extract work.
So we are obsessed with efficiency. How are we going to most affect—how like making the ideal closed system is part of what we’re trying to achieve, right? And then we’re learning that a certain kind of perpetual motion is impossible. So a certain kind of ideally like perfectly closed system is impossible. So like there’s an implicit critique of autopoiesis here, or Ashby and ultrastability. But we don’t need to go there. I think I mentioned it last week as well, but I always have to like flog those guys just a little bit. But then we confuse the nature of reality itself, with this very like exotic artifact that we’ve created. And in fact, there are no closed systems in nature, and there’s no empirical reason. I mean. There are, of course, empirical reasons, but those empirical reasons are themselves filtered through this set of theoretical assumptions to treat the universe itself as a closed system. I mean, you can’t say there are no empirical reasons to do that, but you can say that the data is being framed in a particular way, and it’s underdetermining vis-a-vis. The theory—theory is underdetermined by the data. There are different ways of interpreting that empirical evidence. And I really like what you’re saying about, you know, we’re creating these very extreme conditions.
And of course they might bear some similarity to our best model of the early universe, but partly our best model of the early universe is predicated on us. Looking back so far that we encounter something which we can’t interpret in terms of our models which are derived from, you know, empirical evidence that we do derive from, or kind of against the point that we’re making in general here, but are partly validated, based on things we do have.
Matt Segall: The singularity is the limit case of relativity. If you—yes.
Timothy Jackson: Exactly.
Matt Segall: Everything back to the origin.
Timothy Jackson: So point of convergence and disappearance, as I’d say, yeah. And that’s the case with with ancestral states in general—that they always encounter that sort of singularity. So no, I think it’s a really, really great point that we’re creating. You know, we’ve got a very exotic setup here, and it’s going to give us some insight into. You know what things are made of, and a certain kind of fungibility, like a transformability of states, of arrangements, of structures. And and you know, very important to do that, and often helps us to break down certain assumptions that we have about, you know, fundamental nature of reality. Because we do a Poincaré move. You know, we break it apart, and we find there’s another layer. And then there’s another, you know, and we’re like some of our models, then, should reveal themselves as like emergent limits, so like certain determinism and maybe obsession with symmetries. For example.
It’s not that symmetries are just broken. But symmetries are fundamental. Maybe it’s that like symmetries themselves are emergent consequences of certain relational dynamics. I mean, this would be where I would go further than Anderson, which you mentioned at the beginning, and “More is Different.” And you and I will do that paper, and we’ll talk about this, and it depends what you mean by symmetry. But yeah, no, I like your point there. I wanted to make another really different point. It’s going to be another kind of slightly hard jump to another thing in the Mcquillan. But yeah. So page eight. Now, which is page eight of the PDF.
Oh, there’s a lot of different little notes that I’ve made here. But there’s a note about alienation here, because you’re talking about data science having more—we mentioned this earlier on, you know, where data science would have more sway than the testimony of the subject in the legal context, because one would assume, given this Neoplatonic character of data science, which, he says, makes it hard to constrain—makes it creates structural conditions for specific injustices. Blah blah! But it’s the assumption of neutrality and objectivity that the model is going to know the facts better than the person who was actually, you know, present for them. And of course we all know about the unreliability of witness testimony and blah blah, but he goes down a little bit.
He says, “Inverting the traditional slogan of the disability movement, data science seeks to know quote everything about me without me.” And I’ve got this note here about alienation, which I think is really interesting, because, you know, one of the things that’s going on, of course, now, and even much more since Mcquillan was writing this, is that these, you know, algorithmic processes are, you know, literally used to create models of us, right? And mostly us as consumers, right? So they can more effectively sell things to us and people like us.
But they are—again, there’s this self-fulfilling prophecy aspect. There’s this feedback aspect where they are, you know, predicting what we’re going to buy and actually catalyzing our behavior through making those predictions available to us. But kind of the end trajectory of this sort of process would be what is called a digital twin in personalized medicine, right? So a digital twin, the basic idea would be that you have this high fidelity, physiological model of you, the individual—hence personalized medicine, so that before we even gave you a drug, or before, indeed, you developed a certain disease, we would be able to predict your likelihood of developing that disease with high fidelity, and we’d be able to predict your response to a certain drug. So this is like a digital twin. Well, it is what it’s called a digital twin. And it’s interesting. There’s an interesting relationship there which is not what I wanted to get into. But between these algorithms which are developing models which are used in this consumerist way, like Amazon, or whatever it might be, or you know Cambridge Analytica and manipulating elections, and all of that. And what’s going on in medicine as a potentially beneficial application, although I think we’re very, very long way away from that kind of thing.
But there’s a Simondonian notion of alienation here. So he kind of riffs on Marx’s notion of alienation, where alienation would be, you know, the alienation of the worker from what they’ve produced, or from their power, their capacity from their own labor, like the products of their labor, right? They don’t own it. And this is a big thing that Yanis Varoufakis talks about in the development of his sort of techno-feudalist conception of things where, you know, we are producing all of this content which is being turned into these products, basically, which kind of are the models which both predict and shape consumer behavior. And we’re alienated from that. And so Simondon has this idea extending from Marx’s alienation. Where—and we’ve talked about this, I think before, but where the technical object in general is itself alienated. Here’s like a weird animacy thing—animism type thing going on here. But the technical object contains an alienated chunk of humanity. So like there’s a piece of the soul of the inventor, and maybe the users and all of the people that enable the technical object to do whatever it can do. Given a particular, you know, sociocultural, machinic assemblage.
And then the alienation occurs because we think that the technical object has this kind of like independent existence. And then we even see it as kind of potentially acting against us, you know, acting on us, and we forget that well, all of its agency, all of its creative powers, or whatever are just humanity that we put in it. And I just like this idea. I think that these algorithms, these digital twins, which are like ghosts or something, you know, are nothing but—like in the most extreme sense, they’re nothing but alienated humanity. They are the purely alienated, extracted from our labor, which is like content, creation on, and even just like going into Amazon and buying something in this sense.
But you know they are nothing other than that. I haven’t really developed this line of thinking, yet.
Matt Segall: Hmm. Reminds me a bit of what Marshall McLuhan describes as the amputation that various forms of media technology end up performing on human beings, and that’s related to this concern about thoughtlessness that Mcquillan is discussing drawing on Arendt and her concerns about the banality of evil and how bureaucracy functions. And now it’s not even just human beings in bureaucratic conditions who are thoughtless. We’re just totally—having, we’re increasingly being encouraged to amputate our thinking process to allow these algorithms to make decisions of economic decisions, legal decisions, legislative decisions. And so many Silicon Valley tech, you know, startup owners and leaders and engineers have this idea that our political problems can be resolved through technological means.
Which scares the hell out of me and makes me really, really angry because it’s so disembedded, dehistoricized. And it’s the worst kind of political ideology. And, in fact, that’s driving that.
Timothy Jackson: What I like about what you said just to quickly interject is like, we’ve amputated our thinking process. So on the one hand, we are becoming thoughtless because we are, you know, offloading, cognitive offloading our thought process onto these algorithms. But again, they are nothing other than our amputated thought process, but amputated from continued thought, right? They’re like a structural kind of achievement, a static structure that is the product of our thinking process and kind of an instantiation or manifestation of it, but cut off from that process of living thought.
So they’re like fossilized thought.
Matt Segall: Right? Yeah. And just I think I’ve mentioned this story to you before. But I’m reminded of Marvin Minsky when his book The Emotion Machine came out 2006-ish, you know, visited University of Central Florida, and performed this veil of neutrality that Mcquillan is talking about where you know. I asked him, “If you succeed, which I don’t think you can. But let’s say you succeed in creating a conscious machine. Does it have rights, you know? Does it get paid for its labor?” And he said, “That’s not my problem. You should be asking the politicians,” as if the science and technology of it is neutral, right?
Timothy Jackson: And where does your funding come from, Marvin?
Matt Segall: Jeffrey Epstein, and the Department of Defense, and so on, and so forth.
Timothy Jackson: Yeah, yeah, sorry. Exactly. Yeah.
Matt Segall: Well, I think I think we should wrap up. I have some packing to do in preparation for this conference down in Claremont tomorrow. But this was good. I feel like we stayed focused and really, really dug into this paper in a productive, I hope.
Timothy Jackson: And it’s a really good paper. And I’m looking forward to reading more Mcquillan, including eventually the book. Can I make one final point that I’ve been gesturing towards which is the evolutionary literacy sort of point.
Matt Segall: Sure. Yeah. Go for it.
Timothy Jackson: I mean, I maybe I can bring it out of this thing on the same page here. But you know, talking again about Neoplatonism and so-called, you know, vulgar Platonism is a worldview that is at the same time its own enactment. Again, because the algorithms and the models are directly interacting the world. Unlike previous forms of metaphysics, Neoplatonic data science attenuates the world directly because it is also machinic. Okay. The point that I wanted to make, which I just have underlined here is that it’s also really traditional. It’s traditional, you know, like it’s—and vulgar Platonism is very traditional, right, and I’d say, like vulgar Platonism, has been vulgarized in lots of different ways, you know, and so, like Gnosticism might seem like a really extreme example. But I mean, you know, Christianity has a lot of vulgar Platonism in it, like certain aspects of Christianity. I don’t want to be offensive here. But you know, like the kinds of things that Jung is criticizing in Answer to Job, for example—everyone go and watch out at least the first video on Answer to Job, which I think was really good going to be returning to Jung soon. So that’s a teaser for the future there, but also mathematical Platonism, like we talked about, is a kind of, you know, vulgar Platonism, and I think that something that masquerades, and not consciously—not that the individuals are doing this consciously, but something that presents itself as kind of subversive and like counterintuitive. And, like, you know, science really trades on this, you know. Notion of counterintuitive, you know, like we’re just apes. So why should we expect reality to be like our intuition? But no, it’s more like this hidden layer of, you know, and it’s like, but that’s obviously so intuitive because people have been saying that for like ever, you know, and in some sense that two-world metaphysics is so baked into our worldview in the West. But not only that—this is not counterintuitive, this is not subversive, this is just snapping to the strongest attractor in metaphysical space, basically. And again, it’s the vulgarity of it that’s the issue, because what’s more counterintuitive, it turns out it seems, would be like a genuine, like constructive evolutionary realist thinking.
And so what I wanted to underline here is the difference, again, between the inference of common ancestry and the inference of the Eternal. And so when these papers—and there’s a couple of papers that I’m thinking of hopefully, I’ve got them right here. Shoot there’s one on the—I’m gonna have to dig him up. Sorry because I don’t. Wanna it’ll take me two seconds. Yeah. So “Harnessing the Universal Geometry of Embeddings,” which is Jar et al. This is a preprint. It’s only just come out and “The Platonic Representation Hypothesis” by Huh et al—apologies if I’m mispronouncing the names. But in the “Harnessing the Universal Geometry of Embeddings” they talk about the strong Platonic representation hypothesis. “Our conjecture is as follows: neural networks, trained with the same objective and modality, but with different data and model architectures, converge to a universal latent space, such that the translation between their respective representations can be learned without any pairwise correspondence.” And in “The Platonic Representation Hypothesis”—just to get the hypothesis itself. “We hypothesize that this convergence is driving towards a shared statistical model of reality, akin to Plato’s concept of an ideal reality.” Now, this is like a lot of things we could pick apart here, including, like, you know, Plato’s ideal model of reality being kind of the antithesis of a statistical model, right? But what I wanted to say is that this convergence, right in in the output, in the models derived here, and with different training data and different model architectures.
And so they’re saying that because there’s convergence, there must be something universal behind it. But universal is a term that’s being thrown around very loosely here, and so is Platonic. And I just want to—and this is probably something we’re going to have to keep getting into. But compare this to just some sort of basic evolutionary reasoning.
So the different data sets in the Universal Geometry paper are like, you know, different sets of annotated natural language. And you know, we might just want to point out that all human natural languages are only, you know, 70,000 years old, whatever, right? So, even if they were taking the most distant representatives of human natural language. And they’re not. These are like, you know, English and Spanish and stuff, you know, the shared evolution of like English and a Romance language is just so much. But even if you were taking the most distant representatives, it would be 70,000 years, right. You know the models, they say, are different, the model architectures. But they’re all transformer models. They’re all essentially built on the same fundamental architecture, but doing slightly different things, and it even says with the same—again, I’ll just pull it up so that I’m not misquoting it. “Trained with the same objective and modality.”
So the finality of the models is the same, right? So they’ve got the same goals, basically, and the same modality.
If you were thinking in like evolutionary terms, just like really standard first year, you know, evolutionary biology terms. And I use this example with you before, like the flying squirrel, you know, gliding, marsupial, convergent evolution, and, like, you know, a non-biologist might not be able to tell the difference between those two animals when you, if you like, you know, held them up and showed them, you know, or showed them in flight, right? In gliding. And there’s 160 million years of divergent evolution between these two mammal lineages, right? But basically, that’s when marsupials and placental mammals split 160 million years versus 70,000, or whatever that we’re talking about here. And I mean vastly less than that for the transformer architecture itself, right?
But what we’re actually talking about is a massive degree of homology, even with 160 million years of history, of separation. The squirrel, the flying squirrel, and the sugar glider, say the gliding marsupial, are both exhibiting a very very standard, generic, mammalian body plan, arboreal, mammal body plan, which is shared—like that’s a synapomorphy of mammals. So that’s homology. And then homoplasy which is convergent evolution comes in at a shallower but still, like, you know, pretty deep in comparison, layer of the evolutionary history, which is basically like tall trees and terrestrial predators. So when you have a combination of a shared ancestry and convergent evolution which is basically adaptation to a similar environment and similar selection pressures, you get something that formally—like at the abstract level. Of course not, if you like. Look closely at the tummy of the—you know, if you look at the reproduction or all these other aspects of the organisms, they’re going to be quite different. But at the level of gross anatomy, at the level of a kind of abstract form. You’ve got incredible degree of convergence, then think about like an ichthyosaur and a dolphin, you know, and that’s like—we don’t know exactly, because we don’t know when ichthyosaurs, which are a clade of marine reptiles diverged from the reptiles, which are the ancestors of mammals. But you’re talking like 300 million years. But there’s still heaps of homology. There’s a basic vertebrate body plan. The skeletons are very, very deeply homologous. The product of shared ancestry. The nervous system is very, very much shared ancestry.
But then there is convergence. So they’re both descended from non-marine ancestors. They’ve both gone into the ocean, and so there’s convergence in adaptation to like the viscosity of seawater, so they’re both very streamlined, and they both have fins. They have long snouts, with lots of very sharp, peg-like teeth, because they feed on fish, you know. So again, homology and homoplasy. Think about the vaunted tetrapod limb homology, you know. That’s 400 million years. All right. I don’t want to get silly here, but I will think about like serotonin serotonin evolved. Maybe. I mean at this point the numbers are a bit meaningless because we can’t infer, you know, whether there’s convergence or homology, but you know it’s shared across lineages which we believe might have been split. And again, molecular clock accuracy is very marginal in these time scales. But 1.5 billion years ago, you know, there are, you know, microbial lineages which have serotonin, and like we have serotonin, right? And the point is that there’s so—you need a conception of deep time.
And again the interplay of shared ancestry and convergence to understand why, you know, I don’t know—two things that things of any kind are converging on, or appear to share a similar form. And like this, really, really basic stuff. And in this kind of data science world. But not only there’s this leap from “oh, look! They’re pointing to something the same” to “it must be universal. It must be Platonic or something like that.” And it’s just like. Where is where is the basic scientific literacy? I’m now being a jerk, you know, but it’s like.
Matt Segall: I mean, interdisciplinarity is very difficult, particularly in STEM, you know it’s easier for us humanities, people to, you know, dabble in a variety of things because we don’t need to become—it. But if you’re a, you know data scientist or a molecular biologist, you’re unique, Tim, and your ability to, you know, integrate these various disciplines.
Timothy Jackson: I’m just a dilettante, right. But.
Matt Segall: Well, no, you’re a scientist that is interested in other things, too.
Timothy Jackson: But actually just an interesting point here, then, is that it shows how this cybernetic utopianism, the complexity, science and system science utopianism just hasn’t kind of. I mean, it’s still early days like, especially in evolutionary terms, right? But it just hasn’t been, you know, capitalized on, because the whole point of those frameworks, and they’re profoundly influential and not always in a good way. But is this interdisciplinarity? Is that we will speak a common language, right? I mean, in a sense, that’s the point of a mathematical model, just like it’s the point of—in a different scale. But the scientific name of an organism. You know, we won’t rely on common names. We’ll be able to speak across languages or whatever.
But it seems like, yeah. Anyway. Sorry I’m gonna go on for too long. If I…
Matt Segall: I mean something to just to wrap us up. I mean, I think you and I circle around this issue a lot is that I think there might be a way of splitting the difference here, that there’s a kind of evolutionary Platonism, perhaps, where the difference between inference to a common ancestor, and inference to some kind of eternal form. We might not need to imagine that these are opposed positions, and I think Whitehead’s account of something like eternal objects, as a way of talking about invariance across moments and occasions, and that when you’re talking about something that is inherited, and that, you know, common. The very idea of common ancestry implies that there is something invariant which is inherited.
And you know we don’t want to leap to a two-world dualism. But I think we do need just as we want to really see how much work, metaphysically speaking, variation can do for us. We also don’t want to lose sight of the equally inescapable work that invariance does for us.
Precisely in order to articulate an evolutionary and constructivist ontology, you know,
Timothy Jackson: Yeah, I think there’s some important difference in the sense that there’s a structural equivalence. There’s an abstract equivalence. The operators. The principles are equivalent, but there’s an ontological distinction which may yet make a difference, even though you and I are both going to agree that this is again that point of inferring a primordium, inferring a sufficiently distant ancestral state. We’re only at that point going to be able to fall back on principles, because we’re not going to be able to really, we never can access the concrete.
Matt Segall: Because I think that the problem is, we tend to understand the eternal. We reify the eternal as an origin in time.
Timothy Jackson: Yeah.
Matt Segall: When really it’s the Platonic conception of time as a moving image of eternity and the Whiteheadian conception of concrescence. Where and you know the whole mystical tradition. Going back, you know, to Eckhart and others, is that, like the Eternal is fully present now, you know it’s not the opposite of time. It’s the—it’s the plenitude of time. It’s the superlative form of time. It’s the wholeness of time that’s present in each of the instances of time.
Timothy Jackson: But I think that is consistent with an origin in time. In the set, like in the shared evolutionary ancestry sense, you know, they can be a determinate origin in time. And sometimes we can actually, kind of, broadly speaking, determine that. And sometimes we can’t, depending on how ancient it is really, and how much data we have. Blah blah. But something can be, have an origin in time and then become—and I guess this term is sympaternal, right. But which Jack Bagby, thank you, is like a useful term, which I got from one of your dialogues with him.
But you know, once—and this would even, and we should do morphic resonance at some point like we should do a thing on Rupert’s, on Sheldrake’s idea there, because but it’s Peircean, and you know, but once the thing happens. It is then stabilized and inherited in all subsequent, you know, moments and some things are like—I mean again, I don’t believe in the immortality of the past, right? So I don’t think that has to be the case for everything. But for some things this is the case.
Yeah. And I do think there’s a meaningful difference there. And I think it does come down to—and you and I’ve just been over this so many times, and we have slightly different intuitions, but very similar—it does come. What is your chosen image of the primordium going to be? And why, you know, what are the considerations here? We’re all going to agree that in like immanent evolutionary process there’s a same and different interaction. So something is inherited. But again, how are you going to have this genetic account? How are you going to generalize this onto-genetic account, such that you don’t start with, you know, it doesn’t need to be the subject, but something that’s like a pre-constituted subject. And that, you know. And so then I think we are in the Simondonian sense, you know, able to do this kind of generalization across all of these schemas. And it’s to me, even though Simondon didn’t see it this way. It’s an extension of Darwinian reasoning which is Peircean. So it’s very Peircean in that sense as well. And what what is consistent across all of these things. I mean, there’s always going to be, you know, in any in medias res process. There’s, of course, going to be a role for a certain kind of sameness. But I think, if you look at like the spontaneous origin of like synchronization in a population of like coupled oscillators, or something like that. And look at that as a model of individuation. Then you’ve got this interesting view of a thing which, at the appropriate level of description, and that’s always the caveat, right? You’ve got a field of pure variation. So you’ve got like randomized distribution of oscillators. And then you’ve got a process of resonance, basically, which requires a medium. So you’re not like, you know, there’s not nothing here but because we can’t figure the nothing, you know, but we can get as close as possible with some of these things where that higher order structure which is the emergent synchronic pattern which is the novel individual in a Simondonian sense, but is a formal, becomes a formal constraint that’s not pre-present.
Matt Segall: Yeah, yeah, yeah. I’m tempted to go on. I mean.
Timothy Jackson: Well, this is our. This is our hobby horse, in some.
Matt Segall: Yeah, yeah, we’ll. So we’ll circle back around, I’m sure. Alright, Tim, better wrap it up. Always fun to chat, and let’s talk offline about where we go next.
Timothy Jackson: That’s good man. Great to see you as always.
Matt Segall: Otherwise. Hope you get some sleep tonight.
Timothy Jackson: Yeah, me, too. See you? Bye.
Footnotes2Plato 
What do you think?