Essay published in the seminal journal by the Institute for Future Cryptoeconomics, Vienna.
Disassembling The Truth Machine
A story in two parts
So there is this thing. Some call it a truth machine, others say it is magic. Many have pledged their faith in it as a solution to problems of power, a way to get rid of all the lying, cheating and corrupt politicians once and for all and put the greedy bankers out of their undeserving jobs. This thing is called the blockchain, it was introduced to the world through the cryptocurrency Bitcoin and it has its own way of producing truth. And the following is the story of how it does so…
…followed by a story of how it does not do so.
Part 1:
Functional truth
Let’s begin with the magic. This is a mathematical kind of magic and it is at the core all crypto-currencies, in fact it is the crypto in cryptocurrencies and the backbone of pretty much all data security, authentication, verification, invocation. Yes, invocation. - It is used across several major industries and has also inspired a hidden undercurrent of blockchain superstition and rituals, but more on that later.
The magic of maths. What I am talking about here is cryptographic hashing. It is a process by which some arbitrary data is run through a mathematical process that spits out a short and fixed length string of characters and numbers depending on hashing function used, say for example: 000000000000000018bf622358138392e15833dc0b6b 6d785226002dcfeaa8de. Only an exact data input can produce that specific string, however it is near impossible to reverse this and work out the original data from the string of numbers. (Feel free to try if you don’t believe me.) Now, because any change to the original data would produce a whole different output when hashed, it can be used as proof that a record of data has not been tampered with. And by adding a timestamp, it can also be proven when that data was added. Chain this together and you have a provably secure linear record of events. This peculiar mathematical property has inspired a broad field of research and development since the late 1970s onwards, of hashing functions with different properties and security models (cf. Merkle, 1979; Preneel 2010). It is the basis of things like public key cryptography, merkle trees and digital signatures, all used in cryptocurrencies.
Ok, you might say. So far the magic of cryptographic hashing is helping us to ensure that we have a record of events that we can prove was added at a particular time and has not been tampered with. But how does this help establishing truth about anything? Who decides what is added to the record in the first place? And how do we agree that the event is true?
The answer is everyone.
Potentially.
Well, in a sense anyway.
I will explain this through the architecture of the still-undeniable-archetype-of-cryptocurrencies, Bitcoin, because without some concrete example this might just get too abstract.
Here is how it works:
Bitcoin was introduced to the world as a peer-to-peer decentralized electronic payment system. The intention was to get rid of the need to trust in banks, institutions or any kind of authority to verify transactions and hold the record of accounts. Instead, this would now be done in a decentralized manner, by us, the very people doing the transactions, through a decentralized network of computers. To put it differently, information about who sent what amount to who, and when, would no longer be recorded and enforced by a bank and a legal system, but by nodes in a decentralized network.
So here it is then, the consensus algorithm in Bitcoin, AKA proof-of-work, AKA the truth machine. Take a deep breath and follow me.
-- Transactions are broadcast to the decentralized network.
-- Nodes in the network then group these transactions into “blocks”.
-- The nodes then compete with each other to decide who gets to verify this block of transactions.
-- The competition consists of hashing (remember?) the block of transaction data along with some random number so that the output string fulfills some requirements, it needs to have a certain amount of zeros, like: 0000000000000000a98127ca14a1e9bda07d58fb4093d16436a50fda5d8127b3. Finding such a number is what is called mining in the world of cryptocurrencies.
-- The hashed output is called the proof-of-work and has the characteristics of being difficult to compute and impossible to fake.
-- It is published to the network along with the block of transactions and nonce so that anyone can check that it is indeed a valid proof-of-work (has the right amount of zeros and is indeed the output of the transaction data hashed with the published nonce).
-- Nodes keep trying different nonces until they have found one that when hashed with the transaction data meets that requirement.
-- This block is then considered valid, and any contradicting transactions are considered invalid.
-- The validated block is added to the blockchain, which is a chain of the entire history of validated transactions.
-- Competition then begins again for validating the next block of transactions.
-- Nodes are rewarded for each block that they find with an amount of newly created Bitcoin and/ or fees, as incentive for validating transactions and securing the network.
Take your time and read that again if you need to. It is a highly unusual and dense entanglement of concepts and functions from the fields of cryptography, mathematical probability, game theory and, yes. – A sprinkling of right wing political economy, instrumentalized (Golumbia, 2016). I say “instrumentalized” because these ideas serve a function here that does not necessarily translate directly into a practical right wing economics. (Apologies for any disappointment to critics and fans, but tech and the blockchain are just not that deterministic.)
Let us analyse a bit more what is happening here. The output of this process is indeed a secure and computationally agreed upon record of events (transactions events in this case). But the consensus algorithm here, the “truth machine” produces a very specific kind of “consensus” and “truth”. Let’s say two transactions were contradicting each other. Someone was trying to send the same value token to two different recipients, broadcasting these transactions to different parts of the network. What matters here is that the decentralized network comes to an agreement about which one is valid. It does not matter which one it is, as long as it is just one of them. It does not matter which one is more “true”, more “fair” or anything of the sorts. The truth of what happened is arrived at, and agreed upon, through a competitive hashing process, determined by CPU number-crunching power. Truth in any transcendental sense, or even in any scientific sense is irrelevant and put aside for the sake of arriving at a functional truth of events.
Lets go through this once again. Economic incentives are used to get people to compete with one another in running computations, hashing transaction data with a nonce, until they find one that is valid and thereby validating that block of transactions. And then they start again, competing to verify the next block. The chances of anyone being able to repeatedly determine which transactions are considered true is diminished as the network gets bigger and more decentralized because they would have to control the majority of the network computing the blocks. The “consensus” of the consensus algorithm should therefore not be misunderstood as some sort of negotiated agreement nor authoritative proclamation on the truth of events but rather an incentive driven settlement on events, the version of which is decided on through randomized turns determined by expending CPU power. It’s legitimacy lies not in negotiations, consensus of opinions or some notion of justice or objective truth but in randomness and large numbers generating an operational computational consensus across the network.
I have gone into detail here explaining the consensus-algorithm of the Bitcoin blockchain not to baffle your mind with technicalities but in order to open a discussion about a form of truth that draws its legitimacy from an entirely new source – not god, not science or philosophy, nor democracy, but a functional truth founded on mathematics. I first sensed this new source of truth and legitimacy, and with it an emerging cult of blockchain believers, when at a meet-up some number of years ago an aspiring actor and Bitcoin enthusiast told me “I don’t trust the banks, or believe in any politician – but I do believe in maths.”
Believing in maths instead of humans. We have opened the protocol, now lets do the same with the blockchain belief system. The distrust of humans that runs strong amongst blockchain believers can be traced to at least three different sources. First, on an immediate and perhaps obvious level, is a generalized experience of blatant lying, cheating and stealing by those in power and a sense that financial, legal and political institutions do more to protect the powerful rather than hold them accountable. Secondly, if we go one step deeper, there is the question of language itself. Human language is vague, we can say one thing, mean something else and do yet another. This, for blockchain believers, is another source of distrust. Language is too imprecise, and easy to game (cf. Filippi and Wright, 2015; Wood, 2014; Jentzsch, 2016). Third, and finally, at the deepest level, the level of existence itself, there is a mistrust of our own minds and senses. How do we know that what we are seeing and feeling is really real? How do I know that my mind is not tricking me? Do I even know my own mind? - A gaping uncertainty between our subjective internal sense of selves and our ability to comprehend or access any objective external reality.
The new cult of the blockchain seeks to resolve this anxious paranoid condition by constructing an apparatus that is external to us, encoded in a network that we cannot game and founded on laws that are objective, derived from maths and the universe itself. Humans will always be corrupted, whereas technology is disinterested, does not care about neither power nor even us necessarily. Political and legal institutions are to be replaced by decentralized clusters of automatic smart contracts that will run regardless of whether you want it to or not. Rules will be written in a language that also cannot be gamed, namely code that executes as written. What is written is the execution itself, not just a claim of the execution (cf. Galloway, 2004). And the mathematical phenomena of cryptographic hashing is mobilised as a fixed point of objective certainty and a source of unbreakable security on which this apparatus is built.
Now the funny thing is even cryptographic hashing requires constant development. The security of different hashing algorithms is an area of continuous research and development, needing to be regularly updated (cf. Preneel, 2010). In fact, anyone who looks closer at blockchain applications will see that there is nothing certain, stabile or necessarily true about ANY of the claims associated with it. Sure, the thing works. In the sense that people are making transactions, coding smart contracts and developing new value tokens. But the central claims of blockchain believers in relation to truth and power are never quite fulfilled. This objective outside from which truth of transactions is devolved, the decentralized, immutable and neutral truth machine, is constantly in a state of correction, forking and development, neither quite decentralized, nor immutable or neutral.
The truth machine, while seemingly external to us, is a thing of our creation, and not only that, is also a constantly evolving thing that we keep maintaining, correcting and changing. Here is where the ongoing claims of the blockchain truth machine become important, if still not exactly fulfilled. The promise and belief in the blockchain as a decentralized, neutral truth machine is exactly what mobilizes the very efforts to continue to try and realize it. While certain aspects of the system are the sites of intense rigorous work, tireless maintenance, research, testing and emphatic forking by a broad community of incredible engineers and developers, still, in other aspects, their agency, this ongoing development work, thinking and decision-making, while lauded and admired, is not acknowledged as agency but is instead subsumed into the grand stories of blockchain maximalism. And this creates some serious blind spots. It makes it possible to divorce any assessment of the effects of the system from the claims made of it, because the truth machine is relegated to an objective world external to us, mediating at a scale that is larger than us.
I am now going to disassemble this belief. I am sorry. I do this not for the sake of disenchantment, but because these mathematical phenomena are actually part of a much more wild and rich magic than this. A magic from a world that is alive and full, rather than lonely and paranoid. In the following I am not going to give you a different model, but instead a set of methods. Ways of understanding and thinking about the establishment of truth and ideas of objectivity that does not require an external machine, but allows us to use this truth machine, the blockchain, our machine that we are making and experimenting with, as another tool amongst many. This will also allow us to assess of the effects of this machine and give a better understanding of the agency we have to change it, do something else or get rid of it if we don’t like these. Let me explain.
Part 2:
Finding truth in an indeterminate world
The dividing line that is at the basis of so much confusion in the ways that we relate to machines, markets and maths, the line between the subjective and objective, or humans and technology (subjective, corrupted, vague vs. objective, neutral, precise) or indeed words and things (concepts vs. reality), is a flawed line. The cut is elsewhere. What I am going to do in the next few paragraphs is to explain the thinking of a philosopher and physicist, Barad – who, drawing on such disparate sources as Niels Bohr and Judith Butler, shows us where the cut is, or, rather, how the cut is drawn and redrawn in a continuous process of materialization. By resolving this line, we will hold once again in our grasp a workable method for assessing truth, not in any final way, but in a way that is true to a continuously changing and open-ended world while also accountable and functional.
There is a tendency when countering technological determinism to bring technology back into the realm of the social. But there is a problem with this. Yes, sure, we can see by examining the histories of various inventions and devices that these things we make are defined and designed for specific needs and are shaped by the struggles and contestations over these, by political, social and legal dynamics as much as scientific ones. But it would be wishful thinking on behalf of social scientists and political theorists to claim that these are the only dynamics that affect technological development. We cannot simply make a thing and have it work exactly as we want it to just by saying it is so. There are some things that, well, work, and others that simply don’t whether we will them to or not. Technology is not entirely socially determined, and there is an objective reality, but it is a reality that is not as divorced from our subjective selves and social concepts as assumed.
We will start with Barad’s understanding of the scientific measuring apparatus. (We can keep in our minds our own apparatus as we go through this, piecing back together our disassembled truth machine.) In Bohrian quantum physics, Barad explains, the apparatus is not external to the observed phenomenon, which it would then objectively measure as is usually assumed. Instead, the apparatus is part of the conditions that in fact determine the very characteristics of the observed phenomenon. This was the ground-shifting reading by the quantum physicist Niels Bohr, a reading that seeks to resolve one of the most famous conundrums in quantum physics – whether light is a particle or a wave. For Bohr, the contradicting evidence meant that it has the potential to be either, and how it becomes one or the other is determined in relation to the measuring device. What Barad draws from this, in an updated and expanded notion of Bohr’s thinking, is that all things at the level of the quant is indeterminate. Phenomena become determinate, have particular determined characteristics, in relation to the apparatuses through which they are observed, made visible and made knowable. The apparatus that registers the mark of a phenomenon forms part of the objective conditions for determining the very characteristics of that phenomenon.
This is quite a mouthful to swallow. Reality, truth: indeterminate. These only become determinate as coherent phenomena in relation to an apparatus or body that registers, senses and is marked by it. Now, before the vertigo of relativity throws us entirely off balance, lets define what objectivity can mean for us in such a reworking of observer and observed. Objectivity for Bohr and Barad, instead of being a thing outside ourselves, always just out of the reach of our limited subjective selves, is the ability to accurately describe and recreate the conditions necessary to reproduce the given phenomenon. What this also means is that words (concepts) are not divorced from things (reality) but are part of a process of materialization in the stabilization of phenomena. Descriptions matter, they form part of the materializing apparatuses and their objective account. But this does not mean that we can just invent whatever reality we want by describing and creating concepts determining an otherwise indeterminate reality according to our will. Why? Because we are not alone in this world, we are not the only agencies at work, there are eons of materialization that predate us, and, indeed – not all concepts are effective. This is important. One might say that concepts are effective to the extent that they are able to matter – a play on words used frequently by Barad, in which what matters refers to mattering, and a process of materialization. And this brings us to agency.
The question of agency is closely tied to the question of ethics, responsibility and the possibility for things to be different. But agency is not something that someone has or doesn’t. Agency describes a force that creates a difference in trajectory, mobilizing a power to make things different. An important point here, and the way that Barad expands on Bohr and accounts for the fact that we can’t just decide what works and what doesn’t, is that agency exists as a potential in any thing, every where, not just humans. We are not the only active subjectivities in a container of objects called reality. The world is alive with agency, potential for things to be different, by mobilizing agency, humans or non-humans. What we have here is a huge shift in conceptual grounding that also requires a shift in attention from seemingly insurmountable schizms of the subjective interior and the objective exterior, the separated realms of humans and technology, to an ever-shifting topology of concepts that are part of shaping a material evolution, “technical” and “natural” that giving rise to new concepts.
Does this mean that anything is potentially true? And therefore nothing? Is indeterminacy the same as post-truth? Can we ever know what is true at all in an ever-shifting topology? Post-truth entails a divorce of statement from a material reality, an eclipse of responsibility for statements. The answer to this condition of post-truth is not to go looking for a complete coherence between statement and event, between word and thing, whether in code that executes as written or by inventing a pure mathematical world external to us that retains such coherence. What I have tried to carefully draw out is that the condition of indeterminacy is not the same as post-truth. The form of indeterminacy we have discussed is one that describes the materializing force of statements, making them absolutely responsible and accountable to their effects. We are responsible for the apparatuses we employ and develop exactly because they literally have a materializing effect. Not in a linear manner, but in relation to the other (non-human) agencies at work, in a collaborative effort of materialization.
Our truth machine now no longer exists as a fixed point of reference constructed in an objective clean realm of mathematical certainty. The magic of maths and markets cannot be so easily divorced from us. So what happens to our truth machine? Can we reassemble it? Does it do anything at all? The machine is still there, and it does indeed reconfigure our agency, but not by resolving it once and for all in a higher order of technological neutral execution. No, instead, what we can see now is that we have mobilized non-human phenomena in an apparatus that ties our behaviors together in networks that affect us and the world in new ways, redistributing what is possible and for whom. The exact ways that it does this is for another essay requiring more careful tracing, because the focus here is on the construction of truth. And what we have done in this disassembling of the truth machine is to shift the source of truth from a belief in an absolute objective condition assigned to the magic of maths towards an ever-changing topology where the construction of what is true is an ongoing process of making concepts, ideas and desires material. We are not alone in this ever-evolving materialization. But we are responsible for our part in it, also and especially when we mobilize mathematical phenomena, silicon, fossil fuels and game theoretical into new configurations. _
References
Galloway, Alexander R. 2004. Protocol, How Control Exists After Decentralisation. MIT Press
Preneel, Bart. 2010. “The First 30 Years of Cryptographic Hash Functions and the NIST SHA-3 Competition.” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 5985 LNCS: 1–14.
Wood, Gavin. 2014. Ethereum: A Secure Decentralised Generalised Transaction Ledger. Homestead Draft.