Scan barcode
lucazani11's review against another edition
4.0
Simulation theory, other minds, consciousness in a virtual machine, sim signs and sim blockers, biosims and pure sims, imperfect and perfect simulations, and the Simulation Riposte (and accompanying Philosophical Riposte joke)
trisjdavila's review
informative
medium-paced
3.5
the questions that chalmers raises are, of course, interesting. i do not think that he manages to answer most of them in a satisfactory way, other than “is virtual reality real?”
his arguments start to become convoluted after a while and after too many examples. some of the questions he poses seem to hold much more value than others but all questions are given the same amount of time and thought in this book. i’m glad i read it, it just feels a little messy and as though there’s a lack of some necessary planning.
his arguments start to become convoluted after a while and after too many examples. some of the questions he poses seem to hold much more value than others but all questions are given the same amount of time and thought in this book. i’m glad i read it, it just feels a little messy and as though there’s a lack of some necessary planning.
adnielsen's review
5.0
It took me a few attempts to get through this book but it was worth it in the long run. Here, David Chalmers takes seriously the simulation hypothesis and ponders how this can shed light on classic topics and problems in philosophy from ethics, epistemology, and metaphysics. I think he makes a good case for virtual reality being possibly just as real or meaningful as non-virtual reality. I also particularly liked his incorporation of the Phil Papers survey of beliefs and attitudes of philosophy professors. Overall, I really enjoyed Chalmers’ speculations on “technophilosophy” and I expect to return to this volume in the not too distant future.
ninj's review
5.0
Great book, exploring the viability of whether we could be in a simulation and whether we'd know, along with virtual reality and augmented reality. A lot of updating classic philosophical arguments, such as maxwell's demon, for the technological age. Pulls in a lot of sci-fi classics, including Permutation City, which I read only a month or so earlier.
Makes a convincing case for the reality and meaningfulness of simulated, virtual and augmented realities. Can be dense at times since it's making arguments, especially for similar scenarios, but very rewarding.
Makes a convincing case for the reality and meaningfulness of simulated, virtual and augmented realities. Can be dense at times since it's making arguments, especially for similar scenarios, but very rewarding.
generalheff's review against another edition
3.0
David Chalmers’ work of techno-philosophy Reality+ is at its best when using simulation-based thought experiments to probe the deepest issues in philosophy, such as dualism versus materialism. But it falls down in its claims to refute some forms of scepticism or convince the reader of the reality of objects in Virtual Reality (VR). The arguments used to support these positions are either convoluted or faintly tautological. In short, the book is a thought provoking read but is unlikely to convince the reader of the deeper claims it tries to make. The rest of this review will explore what the book covers and its pros and cons.
What is the book trying to achieve? Chalmers is an engaging, easy-to-read author who tackles a range of classic topics in philosophy – what is real, how do we know anything, the problem of other minds – from a new vantage point: VR and fully-immersive simulations. He uses this standpoint to defend two hypotheses. The simulation hypothesis, which says that “we are and always have been in an artificially designed computer simulation of a world” based in a reality ‘one level up’ from our own; and simulation realism, which says that “If we’re in a simulation, the objects around us are real and not an illusion”.
These are not the only points the book makes but are, in my view, the key ones. The simulation hypothesis could simply be a Matrix-style ‘brain-in-a-vat’ scenario or something more ‘pure’ whereby our brain is entirely created within a simulation. Chalmers argues for this on statistical grounds – essentially saying that, in our reality, we are likely to be able to create such simulations soon. So, what is stopping us being in a simulated world given the vastness of space and the high likelihood some civilisation has beaten us to the punch and begun experimenting with thousands of such simulations already? The ease of spinning up simulations would mean there are far more sim-realities than the non-sim sort, leading to us likely being in such a simulation on the balance of probabilities.
Simulation realism is Chalmers attempt to refute certain forms of scepticism by arguing that the world around us is real regardless of whether we are in a simulation or not. The arguments for this are covered in depth below but a key idea in this is that, were we to discover that the world is underpinned by a digital physics - 1s and 0s (bits) – we would not decry the world as ‘unreal’; we would simply espouse an it-from-bit hypothesis. Why should we claim objects in an immersive simulation are, therefore, unreal despite also being rooted in digital processes? In other words, simulation objects are real.
In addition to these two central claims, the author argues for virtual realism – the idea that even VR of today is real in a certain sense. We are then treated to sections looking at what simulations have to say about ethics and theory of mind. The book closes on a much denser section ‘Foundations’ in which some of the details of simulation realism and its links to structuralism are explored, as well as considerations of whether our consciousnesses could be founded on fleeting ‘dust clouds’ that happen to generate conscious processes instantaneously.
What works in Reality+? At its best, the book leverages the question ‘what if we are in a simulation’ to probe the deepest questions of philosophy. For example, our simulator would be an ersatz god for us. If that simulator, as seems likely, builds and runs many other simulations, they would have little time to take an interest in our world, providing an answer for why ‘god’ so rarely intervenes in our world. The simulator would also likely inject randomness so as not to create a whole host of identical simulations. Given this, it is no wonder that bad things happen, meaning the simulation hypothesis offers an explanation for the presence of evil in the world.
Simulations can be used to explore other topics too. A classic ontological challenge is Cartesian mind-body duality: simulations could offer a concrete illustration of how dualism is objectively true. In a Matrix scenario of hooking brains up to simulations, a sim-Descartes expressing a view that his mind is entirely distinct from the world around him is, literally, correct: his brain sits in reality 1 while his experience of time and space occurs in the simulated world (reality 2). This is backed up with great diagrams to illustrate these ideas; I have never once cited diagrams as a plus (or negative) in a work of philosophy, but they truly stand-out here. Dualism raises epistemological challenges about how the mind substance can come to know anything about the world-substance. The simulation hypothesis offers interesting ways to tackle this too: the simulator, naturally, can set things up in such a way that their subjects can know things by tailoring simulations so the mind-processes and world-processes can interact.
To be clear, beyond the simulation hypothesis itself, Chalmers is not arguing that we are brains in vats or that cartesian dualism is true. Rather, he is using ideas of simulations to probe ideas in philosophy – like the received idea that dualism is wrong. This playful experimentation is where the book shines.
In addition, the book offers a good overview of several trends in modern philosophy through these discussions, covering recent developments in theory of language or the problems of consciousness, like how “identifying a brain mechanism that leads us to classify a stimulus as red doesn’t tell us [about the] conscious experience of redness”. Similarly, the discussions of whether or not chaotic processes – like random fluctuations of dust clouds in the cosmos – could, instantaneously, produce a sentient being like us (effectively just a probabilistic argument based on size and age of the universe) is cogently dissected by the author as well.
What are my issues with the book? At its core, the book is an argument for simulation realism as a means to tackle the threat of scepticism about the external world. This hinges on what I view as sophistical arguments that are not convincing. By this I mean the author arrives at his conclusion that things are real, even if we’re in a simulation, via convoluted syllogisms and subtle adjustments of what we mean by specific words. Perhaps individual steps can be agreed with, but the overall arc of the argument ends up seeming hollow – like a sleight of hand rather than a convincing demonstration.
Chalmers’ discussion of Hilary Putnam is a case in point. The author describes Putnam’s ‘externalist’ approach to language: rather than meaning being ‘all in the head’, Putnam argues that meaning is anchored in the environment. For example, water is what plays the water role – it is not a purely mental construct; it hangs on something out there. Putnam illustrates this with the thought experiment of Earth and Twin Earth. On Earth, water is H2O; on Twin Earth, what is referred to as water has some other chemical structure XYZ. They are, scientifically, different things even if Putnam on Earth and Twin Putnam on Twin Earth do not know this. Putnam argues that despite the use of the same word, and the similarity of the substances, water on Earth and Twin Earth mean different things and, therefore, meaning is anchored in the environment. This could be dubbed a “causal theory of reference”.
The two things Chalmers does next is really the crux of the problem. First, he leverages this causal theory of reference to argue for simulation realism. He begins with Daniel Dennett’s objection to the idea of VR being real that “virtual hurricanes don’t make you wet”. Chalmers argues that, in fact, someone living in a simulation has actually been using the word hurricane to refer to digital (it-from-bit) hurricanes the whole time. A sim person’s word hurricane, per Putnam, has a different (but still causal) referent than a non-sim person’s ‘hurricane’: “If we’re on Sim Earth, then ‘hurricane’, ‘water’ and ‘wetness’ refer to digital entities that people on Earth call virtual hurricanes, virtual water, and virtual wetness.”
The punchline is that the person in the simulation is well within their right to refer to such events as hurricanes, and will go on to experience digital (virtual) wetness. A person saying a (virtual) hurricane doesn’t make you wet is making something of a category mistake –confusing the sim from non-sim setting.
This is problematic. Chalmers is relying on structuralism (water is what plays the water role) to arrive at the idea that (some) objects in virtual worlds are real. But adopting a ‘structuralism’ about language appears to integrate into his argument the very thing the author set out to prove: that things in simulations are real and scepticism is avoided. This is why, for me, the argument rings hollow. It has a hint of circularity about it. At best, it feels a bit cute, like we are simply adding the word virtual in a few places so rendering virtual objects real. A great example of this is the following:
This seems pretty tautological and does not really get us anywhere with establishing ‘reality’ of objects in simulations in my view except in a narrow, quasi-technical sense. It’s also worth pointing out how Chalmers goes on to debunk a similar move of Putnam’s to prove we are not a simulation in a vat. Putnam argued that if I claim I am a brain in a vat, I am making a nonsensical claim similar to the hurricane-wet issue. I don’t have access to the referent of my true brain in my reality so to claim I am a brain-in-a-vat is to make a category mistake. Putnam uses this to debunk Matrix-style brain-in-vat scepticism, but Chalmers argues Putnam is wrong. It is a subtle argument related to which words are truly covered by Putnam’s theory and are “anchored to a specific system”. My issue is that I believe Chalmers is guilty of just the same kind of subtle manoeuvring to refute scepticism and his challenge of Putnam only highlights the weak, dialectical nature of his own approach.
Other arguments are deployed for simulation realism. The author spends considerable time building up the idea of the it-from-bit creation hypothesis and arguing this is equivalent to simulation realism. The argument starts with a “perfect, global, and permanent simulation”. The crucial step is to get ‘its’ – i.e. things in the world of the sim person – to arise from bits. This relies on structuralism in physics: “theories in physics can be boiled down to their structure … a theory in physics is true if this structure is really present in the world. If the structure of atomic physics is really present in the world, for example, then atomic physics is true, and atoms exist”.
This leads to the realist claim: “if we’re in a simulation, most of our ordinary beliefs are true”. But what has Chalmers really done here? He has adopted a structuralist viewpoint about physics stating that structures we identify have reality. By coding the laws of physics from the non-sim world (even approximately) into a simulation, when a sim person refers to an object like a quark they are accurately describing quark structures in the bits of their simulation and so, per structuralism, that quark is real. Put more succinctly: if I build a simulation with a certain set of structures, and sim people identify those structures, they have true beliefs.
Again, I’m honestly not sure what we’ve achieved here. Structuralism, again, appears to be baked into the argument of what we are setting out to prove, undercutting any sense, as Chalmers claims, to have put a dent into the imposing edifice of philosophical scepticism.
There is also an interesting ontological challenge here (at least to my humble eye): Chalmers repeatedly refers to a “perfect” simulation while at the same time using structuralism to impute reality to objects identified in physics. What does it mean to understand our own physics perfectly so as to be able to render it algorithmically? I have not worked out the details of this issue, but it appears to me that the notion of ‘perfect simulation’ clashes with structuralism in some way.
I also believe the very possibility of rendering sentient life within a simulation (and not just philosophical ‘zombies’ like perhaps ChatGPT) is glossed over far too quickly. This and other technical challenges, like the amount of computing power required to create the number of complex, nested simulations that Chalmers needs to motivate his statistical arguments in favour of the simulation hypothesis, are not argued for nearly convincingly enough.
Beyond the core arguments for the simulation hypothesis and simulation realism, some of the claims around VR fall a little flat. Here we are considering not fully immersive, scepticism-inducing simulations but the sorts of VR we have today. Chalmers’ claim is that “when using VR, you are perceiving virtual objects that really exist. They are concrete data structures inside a computer”. My reaction to this “virtual realism” is that it is hard to agree or disagree particularly strongly – does anything hinge on the reality or otherwise of my avatar in VR? Or the ontological status of a concert I attend in VR (“Virtual events really happen. They just happen in virtual reality.”)? This appears to vex the author, as if there is a strong lobby denigrating VR experiences as unreal, which I’m not entirely sure is the case.
Overall, the section on VR is the weakest of the book. It insists on arguing points that are one part uncontroversial to one part inane. Another grating feature of the author’s writing revealed in this section is the incessant spelling out of cases and subcases, as if to deflect hyper-picky readers. I imagine this is something of an occupational hazard for a philosopher, but it gets extremely boring. For example, there is a detailed discussion of deepfakes and the types of ‘sim Hilary / Pizzagate scandal’ and how we could variously describe these as real or otherwise. Elsewhere, there is a tedious enumeration of how various pre-programmed simulations may call into question different aspects of simulation realism (in a discussion leveraging Robert Nozick’s famous ‘experience machine’). None of which is necessary for the points being made at the time.
My last challenge to the book, despite being dressed up in a genuinely interesting techno-philosophical garb, is that it is ultimately not that different to the philosophy of Kant. The German philosopher argued for a metaphysics of a world divided into that which we experience (phenomena, or Chalmers’ ‘its’) sitting above a substrate of unknowable things-in-themselves (noumena, or Chalmers’ ‘bits’). Kant insisted this did not make him a Cartesian sceptic (we can never know the world as it really is) and spent the better part of the Critique of Pure Reason building a complex schema of categories to illustrate how we can have knowledge of the world via his transcendental deduction of the categories of experience. This, according to Kant, is how we can make true statements about the world and not succumb to Humean scepticism in which all our knowledge is only a posteriori.
Chalmers follows straight in Kant’s path. He splits the world into the experience we can know about and the substrate (the simulation) we cannot know about. He even, at one point, introduces Kant’s famous unknowable X that may underlie the bits his simulation hypothesis utilises (“its-from-bits-from-X”). He has the simulator to gel together the world itself with our minds, solving the knowability of the outside world issue just like Kant’s categories. And, like Kant, despite arguing for what would appear to be an extremely sceptical viewpoint – viz. that we cannot know the transcendent thing-in-itself or we are in a simulation – both philosophers believe they have retained reality and avoided the pitfall of scepticism. Kant because he believes he has arrived at the conditions of experience a priori; Chalmers via various arguments in favour of simulation realism (that often pivot around structuralism).
Like Kant, I believe Chalmers fails. Kant’s project was incredibly influential in re-framing metaphysics. His ‘Copernican turn’ pushed philosophers to think about how we, the subject of experience, frame the world and find in it things that we actually put there (like causality). But, on Kant’s terms, his project was a failure. His book, like Chalmers, explicitly set out to put a nail in scepticism. Yet subsequent thinkers like Hegel went out of their way to re-interpret his works to remove the unknowable thing-in-itself, viewing it as a deeply sceptical, unwanted artefact. Kant even added additional material to his first Critique to limit the accusation that his thing-in-itself turned him into a Berkelian idealist, such was his concern at being seen to fall into total idealism (in other words, total scepticism about the external world).
I raise all this for two reasons. Firstly, because Chalmers only draws the parallel with Kant right at the end of the book (“my interpretation of the simulation hypothesis clearly has a Kantian flavour”) whereas it deserves acknowledging much earlier on. Secondly, and more problematically, I believe that having at last stated the parallel, it is extraordinary that such a well-regarded philosopher does not learn the lessons of 200 years of philosophy and attempt to arm himself against the pitfalls into which Kant’s project fell and that threaten his own. Specifically, that he too will be charged with a thoroughgoing scepticism by creating an unknowable thing-in-itself (bits, or bits-from-X) which his simulation realism forever dooms us with (however ‘real’ we may incessantly label the objects around us).
To wrap up, this is an engaging read. It is on far stronger ground when re-interpreting old philosophy in modern technological terms than when it attempts to break new ground. An extremely thought-provoking and worthwhile read, to be sure, but one that, like Kant, fails in its central aim.
What is the book trying to achieve? Chalmers is an engaging, easy-to-read author who tackles a range of classic topics in philosophy – what is real, how do we know anything, the problem of other minds – from a new vantage point: VR and fully-immersive simulations. He uses this standpoint to defend two hypotheses. The simulation hypothesis, which says that “we are and always have been in an artificially designed computer simulation of a world” based in a reality ‘one level up’ from our own; and simulation realism, which says that “If we’re in a simulation, the objects around us are real and not an illusion”.
These are not the only points the book makes but are, in my view, the key ones. The simulation hypothesis could simply be a Matrix-style ‘brain-in-a-vat’ scenario or something more ‘pure’ whereby our brain is entirely created within a simulation. Chalmers argues for this on statistical grounds – essentially saying that, in our reality, we are likely to be able to create such simulations soon. So, what is stopping us being in a simulated world given the vastness of space and the high likelihood some civilisation has beaten us to the punch and begun experimenting with thousands of such simulations already? The ease of spinning up simulations would mean there are far more sim-realities than the non-sim sort, leading to us likely being in such a simulation on the balance of probabilities.
Simulation realism is Chalmers attempt to refute certain forms of scepticism by arguing that the world around us is real regardless of whether we are in a simulation or not. The arguments for this are covered in depth below but a key idea in this is that, were we to discover that the world is underpinned by a digital physics - 1s and 0s (bits) – we would not decry the world as ‘unreal’; we would simply espouse an it-from-bit hypothesis. Why should we claim objects in an immersive simulation are, therefore, unreal despite also being rooted in digital processes? In other words, simulation objects are real.
In addition to these two central claims, the author argues for virtual realism – the idea that even VR of today is real in a certain sense. We are then treated to sections looking at what simulations have to say about ethics and theory of mind. The book closes on a much denser section ‘Foundations’ in which some of the details of simulation realism and its links to structuralism are explored, as well as considerations of whether our consciousnesses could be founded on fleeting ‘dust clouds’ that happen to generate conscious processes instantaneously.
What works in Reality+? At its best, the book leverages the question ‘what if we are in a simulation’ to probe the deepest questions of philosophy. For example, our simulator would be an ersatz god for us. If that simulator, as seems likely, builds and runs many other simulations, they would have little time to take an interest in our world, providing an answer for why ‘god’ so rarely intervenes in our world. The simulator would also likely inject randomness so as not to create a whole host of identical simulations. Given this, it is no wonder that bad things happen, meaning the simulation hypothesis offers an explanation for the presence of evil in the world.
Simulations can be used to explore other topics too. A classic ontological challenge is Cartesian mind-body duality: simulations could offer a concrete illustration of how dualism is objectively true. In a Matrix scenario of hooking brains up to simulations, a sim-Descartes expressing a view that his mind is entirely distinct from the world around him is, literally, correct: his brain sits in reality 1 while his experience of time and space occurs in the simulated world (reality 2). This is backed up with great diagrams to illustrate these ideas; I have never once cited diagrams as a plus (or negative) in a work of philosophy, but they truly stand-out here. Dualism raises epistemological challenges about how the mind substance can come to know anything about the world-substance. The simulation hypothesis offers interesting ways to tackle this too: the simulator, naturally, can set things up in such a way that their subjects can know things by tailoring simulations so the mind-processes and world-processes can interact.
To be clear, beyond the simulation hypothesis itself, Chalmers is not arguing that we are brains in vats or that cartesian dualism is true. Rather, he is using ideas of simulations to probe ideas in philosophy – like the received idea that dualism is wrong. This playful experimentation is where the book shines.
In addition, the book offers a good overview of several trends in modern philosophy through these discussions, covering recent developments in theory of language or the problems of consciousness, like how “identifying a brain mechanism that leads us to classify a stimulus as red doesn’t tell us [about the] conscious experience of redness”. Similarly, the discussions of whether or not chaotic processes – like random fluctuations of dust clouds in the cosmos – could, instantaneously, produce a sentient being like us (effectively just a probabilistic argument based on size and age of the universe) is cogently dissected by the author as well.
What are my issues with the book? At its core, the book is an argument for simulation realism as a means to tackle the threat of scepticism about the external world. This hinges on what I view as sophistical arguments that are not convincing. By this I mean the author arrives at his conclusion that things are real, even if we’re in a simulation, via convoluted syllogisms and subtle adjustments of what we mean by specific words. Perhaps individual steps can be agreed with, but the overall arc of the argument ends up seeming hollow – like a sleight of hand rather than a convincing demonstration.
Chalmers’ discussion of Hilary Putnam is a case in point. The author describes Putnam’s ‘externalist’ approach to language: rather than meaning being ‘all in the head’, Putnam argues that meaning is anchored in the environment. For example, water is what plays the water role – it is not a purely mental construct; it hangs on something out there. Putnam illustrates this with the thought experiment of Earth and Twin Earth. On Earth, water is H2O; on Twin Earth, what is referred to as water has some other chemical structure XYZ. They are, scientifically, different things even if Putnam on Earth and Twin Putnam on Twin Earth do not know this. Putnam argues that despite the use of the same word, and the similarity of the substances, water on Earth and Twin Earth mean different things and, therefore, meaning is anchored in the environment. This could be dubbed a “causal theory of reference”.
The two things Chalmers does next is really the crux of the problem. First, he leverages this causal theory of reference to argue for simulation realism. He begins with Daniel Dennett’s objection to the idea of VR being real that “virtual hurricanes don’t make you wet”. Chalmers argues that, in fact, someone living in a simulation has actually been using the word hurricane to refer to digital (it-from-bit) hurricanes the whole time. A sim person’s word hurricane, per Putnam, has a different (but still causal) referent than a non-sim person’s ‘hurricane’: “If we’re on Sim Earth, then ‘hurricane’, ‘water’ and ‘wetness’ refer to digital entities that people on Earth call virtual hurricanes, virtual water, and virtual wetness.”
The punchline is that the person in the simulation is well within their right to refer to such events as hurricanes, and will go on to experience digital (virtual) wetness. A person saying a (virtual) hurricane doesn’t make you wet is making something of a category mistake –confusing the sim from non-sim setting.
This is problematic. Chalmers is relying on structuralism (water is what plays the water role) to arrive at the idea that (some) objects in virtual worlds are real. But adopting a ‘structuralism’ about language appears to integrate into his argument the very thing the author set out to prove: that things in simulations are real and scepticism is avoided. This is why, for me, the argument rings hollow. It has a hint of circularity about it. At best, it feels a bit cute, like we are simply adding the word virtual in a few places so rendering virtual objects real. A great example of this is the following:
You might object that a creature in a simulation would have many false beliefs. For example, a sim might think, ‘I’m in New York,’ when in fact the simulation is running on a server in Silicon Valley. Is the sim’s belief false this time? No! When the sim says, ‘New York,’ the name doesn’t refer to the unsimulated New York on Earth. It refers to a place on Sim Earth: Sim New York. The sim is indeed in Sim New York … when a sim says ‘in,’ that word means virtually in,’ … so when the sim thinks ‘I am in New York,’ this means that the sim is virtually in Sim New York, which is true.
This seems pretty tautological and does not really get us anywhere with establishing ‘reality’ of objects in simulations in my view except in a narrow, quasi-technical sense. It’s also worth pointing out how Chalmers goes on to debunk a similar move of Putnam’s to prove we are not a simulation in a vat. Putnam argued that if I claim I am a brain in a vat, I am making a nonsensical claim similar to the hurricane-wet issue. I don’t have access to the referent of my true brain in my reality so to claim I am a brain-in-a-vat is to make a category mistake. Putnam uses this to debunk Matrix-style brain-in-vat scepticism, but Chalmers argues Putnam is wrong. It is a subtle argument related to which words are truly covered by Putnam’s theory and are “anchored to a specific system”. My issue is that I believe Chalmers is guilty of just the same kind of subtle manoeuvring to refute scepticism and his challenge of Putnam only highlights the weak, dialectical nature of his own approach.
Other arguments are deployed for simulation realism. The author spends considerable time building up the idea of the it-from-bit creation hypothesis and arguing this is equivalent to simulation realism. The argument starts with a “perfect, global, and permanent simulation”. The crucial step is to get ‘its’ – i.e. things in the world of the sim person – to arise from bits. This relies on structuralism in physics: “theories in physics can be boiled down to their structure … a theory in physics is true if this structure is really present in the world. If the structure of atomic physics is really present in the world, for example, then atomic physics is true, and atoms exist”.
This leads to the realist claim: “if we’re in a simulation, most of our ordinary beliefs are true”. But what has Chalmers really done here? He has adopted a structuralist viewpoint about physics stating that structures we identify have reality. By coding the laws of physics from the non-sim world (even approximately) into a simulation, when a sim person refers to an object like a quark they are accurately describing quark structures in the bits of their simulation and so, per structuralism, that quark is real. Put more succinctly: if I build a simulation with a certain set of structures, and sim people identify those structures, they have true beliefs.
Again, I’m honestly not sure what we’ve achieved here. Structuralism, again, appears to be baked into the argument of what we are setting out to prove, undercutting any sense, as Chalmers claims, to have put a dent into the imposing edifice of philosophical scepticism.
There is also an interesting ontological challenge here (at least to my humble eye): Chalmers repeatedly refers to a “perfect” simulation while at the same time using structuralism to impute reality to objects identified in physics. What does it mean to understand our own physics perfectly so as to be able to render it algorithmically? I have not worked out the details of this issue, but it appears to me that the notion of ‘perfect simulation’ clashes with structuralism in some way.
I also believe the very possibility of rendering sentient life within a simulation (and not just philosophical ‘zombies’ like perhaps ChatGPT) is glossed over far too quickly. This and other technical challenges, like the amount of computing power required to create the number of complex, nested simulations that Chalmers needs to motivate his statistical arguments in favour of the simulation hypothesis, are not argued for nearly convincingly enough.
Beyond the core arguments for the simulation hypothesis and simulation realism, some of the claims around VR fall a little flat. Here we are considering not fully immersive, scepticism-inducing simulations but the sorts of VR we have today. Chalmers’ claim is that “when using VR, you are perceiving virtual objects that really exist. They are concrete data structures inside a computer”. My reaction to this “virtual realism” is that it is hard to agree or disagree particularly strongly – does anything hinge on the reality or otherwise of my avatar in VR? Or the ontological status of a concert I attend in VR (“Virtual events really happen. They just happen in virtual reality.”)? This appears to vex the author, as if there is a strong lobby denigrating VR experiences as unreal, which I’m not entirely sure is the case.
Overall, the section on VR is the weakest of the book. It insists on arguing points that are one part uncontroversial to one part inane. Another grating feature of the author’s writing revealed in this section is the incessant spelling out of cases and subcases, as if to deflect hyper-picky readers. I imagine this is something of an occupational hazard for a philosopher, but it gets extremely boring. For example, there is a detailed discussion of deepfakes and the types of ‘sim Hilary / Pizzagate scandal’ and how we could variously describe these as real or otherwise. Elsewhere, there is a tedious enumeration of how various pre-programmed simulations may call into question different aspects of simulation realism (in a discussion leveraging Robert Nozick’s famous ‘experience machine’). None of which is necessary for the points being made at the time.
My last challenge to the book, despite being dressed up in a genuinely interesting techno-philosophical garb, is that it is ultimately not that different to the philosophy of Kant. The German philosopher argued for a metaphysics of a world divided into that which we experience (phenomena, or Chalmers’ ‘its’) sitting above a substrate of unknowable things-in-themselves (noumena, or Chalmers’ ‘bits’). Kant insisted this did not make him a Cartesian sceptic (we can never know the world as it really is) and spent the better part of the Critique of Pure Reason building a complex schema of categories to illustrate how we can have knowledge of the world via his transcendental deduction of the categories of experience. This, according to Kant, is how we can make true statements about the world and not succumb to Humean scepticism in which all our knowledge is only a posteriori.
Chalmers follows straight in Kant’s path. He splits the world into the experience we can know about and the substrate (the simulation) we cannot know about. He even, at one point, introduces Kant’s famous unknowable X that may underlie the bits his simulation hypothesis utilises (“its-from-bits-from-X”). He has the simulator to gel together the world itself with our minds, solving the knowability of the outside world issue just like Kant’s categories. And, like Kant, despite arguing for what would appear to be an extremely sceptical viewpoint – viz. that we cannot know the transcendent thing-in-itself or we are in a simulation – both philosophers believe they have retained reality and avoided the pitfall of scepticism. Kant because he believes he has arrived at the conditions of experience a priori; Chalmers via various arguments in favour of simulation realism (that often pivot around structuralism).
Like Kant, I believe Chalmers fails. Kant’s project was incredibly influential in re-framing metaphysics. His ‘Copernican turn’ pushed philosophers to think about how we, the subject of experience, frame the world and find in it things that we actually put there (like causality). But, on Kant’s terms, his project was a failure. His book, like Chalmers, explicitly set out to put a nail in scepticism. Yet subsequent thinkers like Hegel went out of their way to re-interpret his works to remove the unknowable thing-in-itself, viewing it as a deeply sceptical, unwanted artefact. Kant even added additional material to his first Critique to limit the accusation that his thing-in-itself turned him into a Berkelian idealist, such was his concern at being seen to fall into total idealism (in other words, total scepticism about the external world).
I raise all this for two reasons. Firstly, because Chalmers only draws the parallel with Kant right at the end of the book (“my interpretation of the simulation hypothesis clearly has a Kantian flavour”) whereas it deserves acknowledging much earlier on. Secondly, and more problematically, I believe that having at last stated the parallel, it is extraordinary that such a well-regarded philosopher does not learn the lessons of 200 years of philosophy and attempt to arm himself against the pitfalls into which Kant’s project fell and that threaten his own. Specifically, that he too will be charged with a thoroughgoing scepticism by creating an unknowable thing-in-itself (bits, or bits-from-X) which his simulation realism forever dooms us with (however ‘real’ we may incessantly label the objects around us).
To wrap up, this is an engaging read. It is on far stronger ground when re-interpreting old philosophy in modern technological terms than when it attempts to break new ground. An extremely thought-provoking and worthwhile read, to be sure, but one that, like Kant, fails in its central aim.
ericness's review
informative
reflective
slow-paced
4.5
This book is an intriguing exploration of a lot of the ideas around virtual reality and artificial intelligence. It's good for the general reader although you may want to skip through some of the detailed arguments. It opened my mind a lot to the idea that in addition to having strong in-person relationships, being part of virtual communities can be a rewarding experience.