Most of the talk about Bitcoin has  centered on its potential as a new form of currency, or on the use of  the underlying technology as a new electronic value transfer platform or  protocol. At the payments industry conference Money20/20  in the autumn of 2013, Bitcoin promoters were declaring that Bitcoin  was “like SMTP [an email protocol] for money.” It promised a universal  protocol allowing different existing payments providers to transfer  value using the Bitcoin blockchain to any other endpoint. At the same  conference, Bitcoin advocates were stickering the exhibit hall, posting  hand-written signs, and passing out leaflets and magazines. By 2014,  however, the promise of the protocol seemed to have come to fruition:  Bitcoin-based payment providers like BitPay and Coinbase purchased  elaborate booths for Money20/20,  and had developed professional collateral to advertise their  services—and, perhaps more importantly, showed their desire for  legitimacy and market dominance alongside traditional payment providers,  such as Western Union, Visa, or PayPal.

About this same time, however, there  were ripples (so to speak) in the Bitcoin universe. New, non-Bitcoin  based startups launched, each a little different in terms of the  underlying technology and promise. A new company called Ripple  put itself forward as an electronic settlement infrastructure, based on  distributed consensus in a communications network rather than a  distributed blockchain database. It would go on to operate almost like  an interbank clearing service, until it was fined by the Financial  Crimes Enforcement Network (FinCEN) of the US Department of Treasury for  operating as an unlicensed money transmitter.[i]  Ripple inherits the ledger technology from Bitcoin but utilizes a  network of trusted parties (“validation” nodes), much like the way  traditional financial services operate, instead of the computationally  expensive and slower mining system of Bitcoin. Additionally, while  Bitcoin is a system of currency, Ripple is a system that transfers debt  obligations, not money. Other start-ups, Ethereum and Eris,  each offered distributed, blockchain-based systems for creating and  running applications without a central server or authority as a control  point. These services promise a way to use the core qualities of  blockchain databases to craft peer-to-peer applications that can be  autonomous and self-executing. Ethereum and Eris also inherit the  blockchain technology from Bitcoin but generalize its use beyond the  exchange of currency, putting software code, not transaction data, on  the blockchain. Unlike traditional software running on a single  computer, Ethereum and Eris are, in a way, “cloud” service providers—but  they avoid the centralization implicit in traditional cloud service  providers. Unlike, say, using Amazon servers to host your software,  these services are decentralized among peers so that no one peer (or  company) can stop service or act maliciously. One of their primary use  targets is the so-called smart contract, a (decentralized) piece of  software capable of enacting legal contracts autonomously, an idea first  proposed by Nick Szabo in 1997. Blockchains, it seemed, were moving from the money space to the law space.

The blockchain, the ledger and the contract

The core qualities new enterprises like  Ethereum exploit all have to do with underlying features of blockchains  as records-keeping devices, and peculiar ones, at that. A blockchain is a  database or ledger that is distributed among all the nodes in the  network running it (at least in theory). Each node has a complete copy  of the entire database (again, at least in theory). Modifications to the  database have to be verified by enough of the other nodes to warrant  that modification’s validity. Bitcoin uses a lottery-like proof of work  system to effect this, but other systems can do it differently.  Regardless, the key characteristics of a blockchain that make it a  special kind of ledger and that are particularly appealing to developers  and proponents are that it is: distributed, decentralized, public or  transparent, time-stamped, persistent, and verifiable.[ii]

Developers using services like Ethereum  want to use these characteristics of blockchains to create distributed  autonomous organizations that can do different kinds of work without the  intercession of intermediaries, central authorities, the state, the  dominance of one individual, or a controlling junta. The decentralized, distributed character of the thing makes that impossible (so long as the  system remains fully decentralized, since in practical terms such  systems are still subject to collusion).  The publicity, verifiability and time-stamped features make the thing  trustworthy to all peers in the system, who rely on their own  contributions to the maintenance and verification of the blockchain to  warrant its truth. Its relative permanence or persistence makes it a new  kind of authoritative record without any central, overarching  authority.

That seems to be an awful lot of work  for a ledger to do. Or is it? Blockchain systems occasion a  reconsideration of two of the central legal devices of modernity: the  ledger and the contract. There is a vast historiography on the role of  accounting in the rise of capitalism and modernity. There is an even  larger archive of history, jurisprudence and philosophy of the contract.  Indeed, the contract has long been considered as foundational to the  Enlightenment as well as incipient to ancient Greek thought. We know of  such systems today because they have persisted in material form, and  have developed into long-standing bureaucratic apparatuses. The ledger  and the contract have worked in tandem to structure and interpellate  human relationships, creating “Man” [sic] itself. These apparatuses  become powerful because they have been placed in certain social  contexts, negotiations of social context that get inscribed within  ledgers and contracts. For all the hype surrounding the blockchain as a  replacement for these essential apparatuses, the blockchain actually  occupies an old role. Indeed, one could argue that the degree to which  blockchains will succeed in materially replacing the ledger and contract  is correlate to the degree to which it hews closely to these ancient  functions.

We are not going to review the vast  literatures about the origins of modernity. Instead, looking at  selections from the histories of accounting and contract technologies  allow us to reflect on what blockchain systems might do—what they might  already be doing!—to these cornerstones of the modern era.

Our examples are not quite chosen at random. Let’s dive in.

1. Look at Moore’s Modern Methods:

Moore’s Modern Methods was founded in 1909 as a manufacturer of  filing and binder supplies for governments and businesses (it has since  become Moores of London,  retaining the binder tradition in custom made leather portfolios and the  like). Maurer first encountered it during his fieldwork in the British  Virgin Islands. Regular ruled ledger sheets made by Moore’s Modern  Methods (hereafter, MMM) had replaced handwritten and typed property  deeds there in the 1970s.[iii]  The effect of this transition to the MMM’s ledger was to cut off room  for maneuver in land disputes. Where prosaic long-form descriptions of  property boundaries or claims left much open to interpretation, MMM’s  ledger forced everything into standardized rows and columns.

Take a look at MMM’s Nimblex System. The reference to “visible”  record-keeping may sound strange to our ears, but in the early 20th  century it referred to newfangled card and paper devices that make it  possible to see at a glance a great deal of data at once, to provide, as  the ad for Acme Visible Records puts it, “’fact-power’ control!”  So-called blind systems, in contrast, would have been stacked or stapled  sales receipts kept in boxes or card-catalog type systems, hiding that  data away.

MMM’s systems made things visible. They also helped structure  decisions. They controlled workflow and thus structured bureaucratic  processes. They also provided a visible time-stamped record of  transactions, which could be cross-referenced for verifiability.

Although they were “visible,” these technologies were private—that  is, they were not open to just anyone, but only to proprietors,  managers, accountants or employees. Their visibility, however, meant  that audit was easy, and they could be made public in case of disputes:  disputes like breach of contract, failure to deliver, failure to pay by a  specified time, and so forth. Those contracts are “outside” the  ledgers—they require different pieces of paper, different authors and  intermediaries, lawyers and the like. But their efficacy depends on what  is going on inside those ledgers. The ledgers did not embed recourse in  themselves, as Ethereum seems to be trying to do (or perhaps eliminate  altogether, since Ethereum contracts execute their terms automatically),  but could be used if legal recourse was sought. That is to say, if  something were to go wrong—if there were a dispute, or a legal claim—one  could bring the visible records out into the open and put them into  evidence before a court or other third party. The remedy to a problem  is, like the contract, outside of the ledger technology. Ethereum seems to build the contract and the remedy in case of its breach inside the technology, as we will discuss below.[iv]

The very earliest ledgers were stored alongside other private and  precious things, even in an actual room or closet that was accessible  only by the male head of the household and his most trusted (male)  confidants (there is some suggestion that the phrase “coming out of the  closet” arises from this context). Because the double-entry ledger was  the most public part of the accounting system it substituted its  prescribed, facing-page, double system of recording transactions for the  literal security of more private books, such as the daybook or journal  stored in a locked room or safe. According to Mary Poovey, it was only by publishing the rules  of the system of accounting that people would come to believe it  accurately reflected the financial reality. Printing accounting manuals  promoted an internally consistent, reliable standard.

The publication of accounting manuals transported the system of  management from the private closet to a public space, and in so doing  became a critical apparatus for business and helped create modernity  itself. In the early days of the development of double-entry ledger  systems their publication became a vehicle for promoting emerging  mercantile power against the King’s traditional rule. Poovey argues that  the social function of double-entry bookkeeping, as an “apologist for  mercantile honesty” (37), coincided with the appearance of printed books  about it. Such accounting manuals prescribed order to accounts by  prominently depicting balances on facing pages, for every credit on the  one page there is a corresponding debit on the other. The ledgers were  internally accurate according to prescribed rules, but their interface  with the outside world was created whole cloth to ensure that the  ledgers would balance. Interaction with the outside world necessitated  creating balance in a way that might seem, from the outside, erroneous  or even fraudulent. The ledgers substituted formal precision for  accuracy to the external world. A consequence of this fabrication is  that it was difficult to actually determine financial standing. We might  accuse the system of being ethically bankrupt, but such an accusation  is to buy into the form of precision promulgated by the system itself,  then turn it back on to itself from the outside, and expect or demand it  to entertain such realities. This is a category mistake: expecting the  internal order of the books to actually occur in the flesh and blood  world is to put the cart before the horse. In the melee of business,  visually ordered accounts suggested an orderly world, making  believe that the honesty and virtue of business was as plain and  transparent to anyone who cared to inspect the ledger.

Modern systems like the Nimblex also came with manuals. MMM’s 1934  catalog lists 19 instructional publications, with titles like “Selling  Costs and Records,” and leaflets, such as:

Leaflet No. 125—a full explanation of the Duplicate Statement System for the use of traders who render detailed accounts to their customers, weekly, monthly, or quarterly. A real time saver.

Ledgers existed in a system of books, producing social effects  that exceeded transcription and calculation. Taken together, these  books established a mode of government, which William Petty (1623- 1687)  called “political arithmetic.” Petty’s political arithmetic was based  on the surety of “number, weight, and method” so as to compel assent to  the system. The simple mathematics used in the ledger transformed  abstract representations into usable facts for governance.

2. The Compte Rendu:

The Compte Rendu was an accounting of all the expenditures of the  French state created just prior to the Revolution. Published in 1781 by  the minister of finance, Jacques Necker, it presented a rosy picture of  the state of French public finance. This strengthened the position of  the King, as well as Necker’s own political fortunes. The Compte Rendu  produced, in double-entry fashion, all of the revenues and expenditures  of the royal house. According to Jacob Soll,  this both demystified the sovereign, rendering it a “shocking set of  numbers” (140), and also allowed Necker to launch a careful critique:  although his figures showed an overall budget surplus, it also revealed  exactly how much the royal house was spending on frivolous things. Soll  shows how it thereby sowed seeds of revolution.

The Compte Rendu became a best seller, selling 100,000 copies in 1781  and was translated for distribution outside of France. It was, Soll  remarks, one of the most successful works of all time.

Why? Because it rendered things public.[v]  If the Nimblex permitted a visible accounting system that was  nonetheless private, only to be brought out in cases of dispute, the  Compte Rendu brought visibility to the masses. It thus brought the  appearance of accountability—and a very public accountability. The  suggestion is that nothing is kept secret, even if the system is only a  shadow of the transparency and accountability it promotes. And more  retrospective than prospective, it did not guide action so much as  report on past action, showing a record of prior decisions. Its  circulation throughout all of France—and indeed the world—contributed to  a political restructuring in the form of a revolution.

3. Mist, the Ethereum client prototype

Mist is a prototype of the end-user client for Ethereum (the final  client may look and work differently). Mist provides a place to browse  the Ethereum community’s collection of code: to join, invoke, and manage  crypto-assets. The client enables the user to manage multiple  identities and multiple sources of funds (just like MMM, the source of  funds is visible but private). And like the publication of accounting  manuals, the browser sets in place a standard to create a reliable,  trusted set of ledgers (or code blocks running on a blockchain). The  same suggestions of trust, honesty, and contractual rationality we saw  in the standardized rows and columns of MMM and the Compte Rendu are  here manifested through multiple user interface features: the user  cross-references the hash value[i]  of a contract to confirm that the code has not been altered (unlike the  ‘blind’ systems of shrink-wrapped software), a scalar ‘trust value’ is  assigned, and a corresponding alert icon establish the highly visible  symbols of trust.

Let’s follow developer Alex Van de Sande’s example  of an Ethereum marriage contract. Van de Sande reimagines contractual  fine print, since “everyone knows” he suggests, that the fine print “is  to protect the… producer, not you.” In its place, legal analysis of  contractual fine print is transformed into a code audit (prompting some lawyers  to speculate about a need for new legal skills), which readjusts the  dynamic of power from the state or corporation to a liberal subject.  Like MMM, the suggestion is that nothing is left secret—all is visible  for evaluation and auditing. Fact-power control! The problem with this  view, however, is that the code for the contract is as likely to hide  secret clauses or tricks as it is to be meaningfully transparent. When  Szabo originally introduced the concept of smart contracts he did not  see “smart fine print” as a benefit of the system. It was rather a  liability. To overcome opaque and complicated code Szabo suggests that  “transaction semantics” in smart contracts need good visual metaphors to  aid in their interpretation, adding to their transparency.

Mist is the simple front-end to the complicated Ethereum system, and  is an important part in the developers’ goal to democratize contracts.  By providing powerful tools for peer-to-peer contracts the developers  seek to displace transactional costs from fees and the cost and time  associated with hiring legal professionals. Instead of requiring an  official to verify signatures, Mist applies a cryptographic public key  from a corresponding private key associated with a pseudonym. In doing  so Mist is carrying on the tradition of “dematerializing” notaries and  other legal professionals, a tradition that Jean-François Blanchette  (2012) previously discussed in the context of the French government’s  move to establish a notarial public-key infrastructure back in the late  1990s. In this case the French government accepted the cryptographic key  as a representation of the notarial power ultimately invested in the human  notary, who was responsible for actually verifying the contract. The  apparatus of the state, the legal professionals, and the rights and  duties all still existed, but the paper and pen were no longer needed.  Ethereum goes much further than Blanchette’s example, as it replaces the  cryptographic doppelganger with a fully dematerialized verification  method that can be conducted directly by peers without the interference  of legal apparatuses.

In the example of an Ethereum marriage contract, the contractual  apparatus is replaced by the Mist browser. Once the parties have decided  to get married they invoke the appropriate code block and work across a  series of horizontal pages to construct the steps of the marriage  contract. In working from page to page (instead of scrolling to the  bottom of a long page, as is typical in most online fine print) the Mist  browser forces structure on the contract parties. This is a linear  method of interaction that is used to slow the user down so that she may  think carefully about each step, but it sacrifices easy reference from  one part to the next. The most important part of the marriage contract  is the Serpent code that will run on the blockchain.[ii]  The developers make a big deal about the Serpent code being “easy to  read.” In the screenshot above a portion of the sample code reads:

MaxWithDrawDaily = 0.01MaxDurationOfInactivity = 365MinAgeOfMaturity = 12if msg.data[0] == 0:id = msg.data[1] * 2 ^ 128if contract.storage[id]:return(0)contract.storage[id] = msg.data[2]contract.storage[id + 1] = msg.data[3]contract.storage[id + 2] = block.timestamp + msg.data[4]contract.storage[id + 3] = id + 5return(msg.data[1])

Your typical traditional marriage contract, on the other hand, is  authorized by state or religious officials permitting the participants  to join in an “honorable state of matrimony” (or something along these  lines), and it is date-stamped, and likely registered with a government  entity. It is also not usually solely concerned with the disposition of  collective property as this one seems to be—although that, of course, is  deeply embedded in its history. The traditional performance of saying  “I do” in marriage is, in the Ethereum contract, we suppose, the  return() callback invoked in the Serpent code. The dynamics of contract  execution in Mist are very different from the familiar ones embedded in  existing social contexts. The challenge facing the construction of the  Mist browser is to somehow replicate these complicated social forces  while also enabling new functionality—functionality that currently  exists in a software development paradigm and is foreign to social  forces implicit in the history of contracts.

By signing a piece of paper in the eyes of witnesses and a  legally-authorized officiant, and filing the paper with a government  organization, the marriage contract becomes a record alongside others.  The Ethereum blockchain takes on this role as record-keeper, but  substitutes a decentralized database for a state archive. So long as the  Ethereum community maintains the blockchain there is persistence and  preservation, but on the flip-side this means Ethereum contracts are  susceptible to invalidity through obsolescence and boredom. If the  electronic network were shut off, or if everyone moved on to a new  system, there is no paper-based backup archiving the existence (or  execution) of these contracts. This isn’t necessarily as scary as it  sounds, though. Many researchers now realize that human sustenance is  necessary for long-term preservation of working artefacts. Even the 10,000 Year Clock,  shrouded in a mountainside, expects (although does not require) the  occasional visitor to wind the bells. Software preservation necessarily  requires continuous format updating, and greatly benefits from a healthy  cadre of activist-preservers who may alter or even “remix” the software  to keep up with the times or to form “living” derivations as Takhteyev & DuPont  explain. Sticking a boxed copy of software in a display case is not a  solution to long-term preservation, and we need to realize that human  use is probably our best chance for preserving complex systems of  software.

The rub, however, is that high technology is famously faddish, so  whether the network of miners will keep your Ethereum marriage contract  as long as your love remains is an open question. Furthermore, because  blockchain technologies are fundamentally cryptographic in nature, they  have an additional preservation challenge. Cryptography is brittle: if  even a single bit is changed (or “rots”) the hash function no longer  precisely refers to the contract, leaving only a nearly-impossible  mathematical needle-in-the-haystack search as redress (formally, “code  cracking”). These errors are mitigated somewhat by the distributed  nature of the blockchain, but errors propagating across the network and  simple coding mistakes (as we saw with Bitcoin  in 2013) remain realistic concerns. Since redress is built in to  Ethereum (or, really, doesn’t exist at all since no breach of contract  is even technologically possible here), broken code results in a  formally invalid contract. In the eyes of the Ethereum blockchain,  divorce may become programmatic and accidental.

Ledgers, Contracts, and Incentives

It does not take that much effort to see ledgers and contracts as  part of the same modern assemblage. The latter has been dependent on the  former, so perhaps we should not be surprised to see them popping up  together in blockchain universes. Contracts between business parties  rest on ledgers like title records, bank statements, lists of  shareholders or ownership stakes, inventories, and the like, and depend  on time-stamped verification of those records. The first difference  between traditional ledgers and Ethereum is that Ethereum builds other  things into the ledger. For the Nimblex and the Compte Rendu, those  other things—contracts, revolutions—take place outside of the  ledger. And those other things depend on other people to help create and  execute them. The accountant is not doing all the work; there is also  the lawyer and, possibly, the judge or arbitrator.

The second difference is that the persistence and verifiability of  the Nimblex and the Compte Rendu were based on physical technologies and  operations of recording. Now, we are not just drawing a distinction  between the physical and the digital.[viii]  We are drawing a distinction in terms of the effort involved in  maintaining the ledgers. Presumably, if I use one of Moore’s Modern  Methods, or if I am the King of France, I pay my accountants to keep up  the books. But if they quit or die, the materiality of the record  endures, ready to be picked up by the next accountant, or filed away  somewhere. Even if the people go away, the records still maintain their  persistence and verifiability.

What about a blockchain? Earlier we referred to boredom. Without a  community of computers running the protocol and engaging in transaction  verification, the system stops working. Things can start to get out of  control. Multiple chains of transactions can start to grow and the  authenticity of the now-multiple records is thrown into doubt. The  physical ledgers we have been discussing solve this problem with paid  employees and paper. Bitcoin solves it with “mining,” the  incentivization of transaction verification by assigning parts of the  ledger to miners who, competing with each other, win the proof-of-work  lottery. Incentivization is critical to ensure that miners do not grow  bored and stop mining, thereby failing to provide the essential  verification mechanism. Without a network of miners the blockchain  itself may even cease to exist, and with no blockchain there is no  ledger and its history of transactions. Much as with old-time physical  world mining: you’re all digging, digging, but sometimes, eureka!, you  strike gold. That is what keeps you digging.

The introduction of contracts is potentially changing the blockchain  ecosystem. By shifting from cryptocurrency to cryptocontract, it changes  the incentive structure for ledger verification. Instead of getting  money for verifying transactions, as with Bitcoin, you will now get  “ether” (in  Ethereum), the fuel of the system, which is also described  as a “token,” borrowing language from the Colored Coins Whitepaper.  Tokens are the reward for doing the work of ledger verification. One  does not strike gold with tokens, and therefore tokens provide fewer  incentives for verifying transactions. We think we are seeing a  discursive shift from mining to tokens. We also think that this is  significant.

Szabo and Buterin  both model the token basis of smart contracts in blockchains on  game-theoretic “Schelling” or focal points. Focal points are a way of  understanding how two parties can come to mutual agreement in an  information-poor environment. Schelling offers the example of two people  in New York who need to meet in an undetermined location without having  prior communication or the ability to ‘contract’ a mutually agreed  location. On game theory assumptions, each party will pick Grand Central  Station (or, at least, will sometimes do so), because in a world in  which many possibilities are equally likely, humans might look for  patterns or unusual focal points of reference—the assumption being that  if I think Grand Central Station is a good meeting spot the other party  may think the same. Buterin offers the example of a fabricated incentive  structure where two people will earn money if they are able to pick the  same number from a list of numbers without communication or  collaboration. People will, according to the theory, look for something—anything—that  makes one number stand out—perhaps an even number, or a number with  many zeros, or a well-known lucky number. Sometimes, so it goes, both  parties will choose “10” and by the miracle of focal points the parties  will have contracted to their mutual benefit without prior  communication. In the less-fabricated case of meeting in New York, there  are actually a great many focal points—Times Square, One World Trade  Center, Brooklyn Bridge, New York Public Library, who knows where else!?  Thus, the chance of our two parties meeting in this example is low. The  situation is similar to contract negotiations between a workers’ union  and a corporation, where both sides are recalcitrant and there are a lot  of issues and too few shared interests and ideas. On the other hand,  Szabo suggests that some focal points are actually quite obvious, or  “hard” in his parlance—the price tag on a retail good is not usually an  invitation to start haggling. In some cultures the stated price is such  an invitation, but in ours the stated price functions as a hard focal  point that greatly speeds contractual negotiations. No pay, no play, as  they say.

So, the development of cryptocontracts such as Ethereum has  precipitated a number of significant changes. Cryptocontracts tend to  build social and functional properties within the system, whereas  traditional contracts require a cadre of individuals to perform these  things outside the contract. Ethereum has also shifted the incentive  structure of existing blockchain technologies from mining to tokens, and  consequently has introduced concerns for verification and longevity.  Concerns about alternatives to the trustless system offered by Bitcoin  have not been lost on the developers of these newer systems; Ripple has  an answer, Ethereum has an answer, but only time will tell. The shift to  tokens is not without its benefits. Speed and efficiency are immediate  technical benefits. Ethereum’s use of tokens takes advantage of a model  of contracts that employ focal points. Focal points, the theory says,  enable mutually-beneficial contracts in a relatively trustless but  incentivized environment.

An interesting thing has happened along the way—or has it?

Seen one way, self-executing smart contracts seem to miss the whole  point of contracts: that, like promises, they are made to be broken.  That is to say, contracts only really get interesting in their initial  formation and in their potential for breach. Ethereum seeks to put  boundaries around uncertainty to the point of snuffing it out.  Contracts, by contrast, are all about managing uncertainty. And other  work gets done alongside that management. Emile Durkheim’s classic discussion  of the non-contractual basis of contract drew attention to all the  social effort backgrounded by the modern belief in the purity of  contract. Durkheim wrote that “facts which are beyond volition” (207)  are always involved in contracts. Further,

it is in the play of these relations themselves that social action  makes itself felt. For everything in the contract is not contractual.  The only engagements which deserve this name are those which have been  desired by the individuals and which have no other origin except in this  manifestation of free will. Inversely, every obligation which has not  been mutually consented to has nothing contractual about it. But  wherever a contract exists, it is submitted to regulation which is the  work of society and not that of individuals, and which becomes ever more  voluminous and more complicated (211).

Stewart Macaulay, writing  in the early 1960s, explained how these ever more voluminous and  complicated social regulations of contracts, these non-contractual  elements, were not only helping businessmen manage uncertainty but also  maintaining good business relationships. The production of an abstract  contract as imagined by the legal scholars against whom he was  writing—and perhaps to be instantiated in Ethereum—missed the point that  the market of contract production as such, with its atomistic  individuals meeting in a zone of free flowing information available to  all, simply does not exist. The real bazaar is nothing like the  hypothetical bazaar where everyone has access to all the same  information and the price mechanism can operate as the textbooks say it  should. People in the bazaar do not search for the best product at the  best price by assuming an omniscient position and taking in all the  available information—nor do they use their feet and walk to every stall  for the same effect. Instead, they conduct “intensive” not “extensive”  searches, and make their best choice. Clifford Geertz explained:

Search is primarily intensive because the sort of information one  needs most cannot be acquired by asking a handful of index questions of a  large number of people, but only by asking a large number of diagnostic  questions of a handful of people (32).

From this perspective, smart contracts are not contracts at all  because there is no possibility of uncertainty in their execution and  thus no compliance; strictly speaking they are just automaticity created  by the verification game. Buterin  even admits as much about Ethereum: “I now regret calling the objects  in Ethereum ‘contracts,’ as you’re meant to think of them as arbitrary  programs and not smart contracts specifically”.

Seen another way, however, the code itself demands a kind of  non-contractual basis of contract. As we saw, one of the ways in which  blockchain-based contract systems work is via tokens or points rather  than money earned through mining. Schelling points are based on  expectations of what everyone else expects everyone else to do. The  verification game itself rests on a kind of background common sense that  sounds a lot like 20th century legal realist Karl Llewellyn’s community of merchants all acting “reasonably” (see Maurer and Richland 2012).  So, in this sense, smart contracts are just like other contracts in  that they ultimately rest on a vague but widely shared (dare we say  distributed?) common sense. Which is itself a cultural system or, as  Durkheim had it, the “regulation which is the work of society”. Smart  contracts afford us an opportunity to reflect on the forces required to  ensure their active constitution—in part these are technical  affordances, surely, but importantly the social and psychological  systems that make up the core of what smart contracts are dictate what  they are able to do.

If there are Schelling points in Ethereum contracts, written in  complicated Serpent code, they are “hard” points. Hard focal points, you  will recall, are like the price tag on a retail good that greatly  reduces the negotiation required for exchange. In fact, the focal points  in Ethereum are so hard as to eradicate the very idea of being modeled  as a focal point at all. The Serpent code does not expedite contract  negotiation, but rather replaces the difficult social and psychological  work of contracting with self-executing code. Schelling points do not  require prior arrangement to arrive at a mutually satisfactory  conclusion, but they do require a certain amount of slippage or slop in  the world (the chaos of New York in Schelling’s original example). In  the case of smart contracts all the messy business and legal apparatuses  have been pushed out of the way: discourse has become binary, and the  conditions of exchange have been rendered into objects. The sacrifice of  exchange that Georg Simmel (2004) discusses  with respect to money—the “interposition of man and his object of  demand”—is offloaded to mediating code and is hardly felt, if at all  (and is certainly not a focal experience). In this case, Lawrence  Lessig’s simple and pithy phrase was right: “code is law” in that code replaces  law. If we must insist on a game-theory model to capture the Ethereum  wedding contract discussed above it isn’t Schelling points, it is the  prisoner’s dilemma.

Second, the verification method is based on people incentivized by  the prospect of gaining tokens. This is the shift from mining to tokens  we mentioned earlier. We are still not sure what to make of this, and  time will tell with systems like Ethereum whether anyone will do the  work of verification for tokens in themselves. Right now, Ethereum’s  ether will be convertible into Bitcoin and sold on exchanges, thus  entering into the currency space. But must it? And if it does not, that  is to say, if it is not warranted by an economic incentive (nota bene,  like common sense, a cultural system), will anyone still “play?” And,  indeed, there is an element of gamification here. It is also Applied  Behavior Analysis—the use of rewards, often tokens, to shape  behavior—here taken to a new level, to run a supposedly self-executing  system.

Again, is this new? Surely, there were economic incentives in the  maintenance of the tools of Moore’s Modern Methods, or the creation and  use of the Compte Rendu. But anyone who has served as a clerk knows the  simple joy of adding up numbers and finding that the account balances,  or the reward that comes from the successful deployment of office filing  supplies. Where some might decry the bloodlessness of smart contracts  embedded on a blockchain, the reduction to zero of degrees of freedom in  self-executing law-like services, we wonder about these simple  pleasures of non-economic rewards to create the satisfaction necessary  to keep such a system going… indefinitely? Well, at least for as long as  the people and machines running such systems want to imagine themselves  useful by doing so. The whole thing starts to seem less like the  science fiction world of distributed autonomous agents and  self-enforcing agreements made by machines, and more like a game of Go.  That might not be a bad thing. So long as it does not turn into  Monopoly.