OK so you’re talking about encoding rules. As I said, we already do this. Which is not to say it’s trivial, it’s definitely not, but it’s not problematic at the fundamental philosophical level.
I’m flattered you took the time to read the article, but I’m sorry I mentioned my work, because I do not have the time/energy to discuss these issues at present, and though related, they’re well off topic IMO, and as I already said I see this sort of thing as being unsuited to forum discussion anyway.
However, because it jumps out at me, and in case you’re sufficiently interested to look at it again, I’ll just mention that you misunderstood “intentional”, which as explained at the beginning of section 2 is ‘a technical term that means not “deliberate,” but “taking an object,” rather like grammatical transitivity.’ This is a very important concept in philosophy of mind, see any of Daniel Dennett’s many books on the subject. (Anyone interested in such areas cannot afford to neglect Dennett, IMO.)
I understand you cannot explain everything here but I am sorry what part did I misunderstand, or did you misunderstand my arguments perhaps?
I do not see any difference from this understanding and from my argument that you can still indirectly infer something about such intentionality. While not itself in the physical domain (I would argue only until we truly understand tought) we can still infer something about it through physical work and the tools we use for it. This is a very common method in psychology to infer something about our mind. And I am aware of your argument that it is not possible to infer since it is a complex system (beliefs, desires etc) but my argument is that it is exactly these type of complex system we now finally can understand more about because of big data and machine learning discovering patterns we could not possibly have hoped to understand on our own.
And in any case I also argue it is not needed to understand in what I am looking at. As you also point out.
Also thanks for the discussion. It brought up some memories from my psychology studies and a lot of things that is useful to understand from a philosophical standpoint when looking at this from as many angles as possible that is important. I am mostly aware in the two domains of law and psychology and probably some of the finer points eludes me when it comes to the philosophy but if you find any weak spots in my arguments that’s great as it is a way to move forward 
I appreciate your willingness to dig into all this stuff. Unfortunately, I can’t match it, as I said before I don’t have the time/energy and it’s off-topic and all the other stuff I said before. Good luck!
1 Like
Allright appriciate the time you did take Robjf. For sure. I have said it before I think this is possible something that can make the world a better place for my kids. I might not be the one to develop it or move it forward but if I can raise awareness on the issues and bring my views on it to others perhaps one day we can reach the next level from knowing to applying!
2 Likes
Only one article from me today as I am enjoying xmas with kids 
First step is knowing then next step will be applying. But we are not yet there so working more on the knowing…
Cardano community reads Computational Law Article 9
The Representation of Legal Contracts an article by Aspassia Daskalopulu and Marek Sergot ( https://arxiv.org/ftp/cs/papers/0106/0106005.pdf ) continues the work on logic based tools to represent legal contracts with a focus on the more long term contractual aspects like formation of contracts and follow up in performance criteria’s.
First the authors examine a priori art on artifical intelligence and legal contracts and also the economical incentives to create such conctracts as with for example ALDUS project from 1992 that concluded that such contracts where not economically viable to create digitally. The authors however argue that the contracts where too simple (SALE OF GOODS) (NOTE: I also believe that simple contracts can be a great benefit to be automated if in great volumes. Just imagine an invoice template in CNL that can be automatically translated to any language the users of the service wants and change tax automatically based on any national law of the user.)
The authors then discuss what exactly a legal contract is (both a binding agreement and also a document with said agreement.) They then make the distinction between the formation of a contract and the ongoing performance of a contract (Note: This seems very useful as a perspective in Contracts over time theory.)
Formation is then divided into determining if a contract is legally binding, enable specification of requirements, assist the drafters in making the finalized contract. Performance is divided into advertising and reminding the parties what needs to be done and when (Note: For me this confirms that a legal system also needs some form of confirmation system or verification system) and monitors the parties compliance with the agreements and in some case violations. (Note it seems useful to split contract formation and contract performance into two different SOA’s - see article 8 - as they require different things. Contract formation can be done visually and quick and with little verification requirements apart from what can already be confirmed from Cardano itself (identity for example), while for example continued performance of say selling x amount of coffee beans and receiving y amount of coffiana (Cardano sidechain example) will require such things as 3rd party verification with IoT mobiles uploading pictures of said transaction or automated weights confirming as beans are weighted onboard ships or with a Wireless chip following the cargo and confirming the transfer.) The authors discuss that simple transaction where a simple exchange of goods happens does not need a contract. For me this is not a clear cut difference and even the simplest transaction could be useful to have as a contract but a very simple contract.
The authors then discuss long term contracts exemplified with 200-300 pages and running for 20-25 years with reviews every 5 years or so. These type of contracts tend to try to be all encompassing and having for example force majeure clauses. The author argues that the size and complexity of such contracts would be an area well suited for electronic assistance in contract formation as these contracts tends to cover a large portion of the project costs.
The authors then discuss that you need domain expertise in contracts when it comes to indirect information while expressively stated information in contracts are easier to interpret. They also discuss the semantics of what exactly is a contract and what is a document and what does it contain. The authors look at contractual provisions and what such provisions could be:
They also note it is not an exhaustive list or a formal definition but argue it serves as an example of different forms a contract needs to capture. Then a very interesting (to me) discussion about abstracting what a contract is follows:
Note for me this also suggest a layered approach with visualization is a great tool for contracts. We can abstract and when you click on an icon you can go to the next layer and this will allow humans to quickly read a lot of data. I believe big data visualization is a huge field as for example how Cardano transaction can be visualized and we as humans can grasp big data better when it is analyzed and presented to us in a visual way.
The authors then discuss what goes into legally binding agreements - was it made during duress etc- (Note I think a future thing here will be if an offer is made by a hostile program - a hacking program) They authors believe these type of things are a problem of classification and will need tools to classify - for example actions. They then discuss the forming of the contract itself how you will need tools for this during contract negotiations. (Note for me visual contracts that can be negotiatied back and fourth on the blockchain seems very well suited as a tool for this.) and that there is a need for terms to be discovered if are in conflict or even possible (Article 8 i posted discusses a tool for doing this.) But interesting they also look into the human aspect of how humans will try to reach the best possible outcome (note: game theory seems an appropriate field for this) and give some literature reference to this I should possible read up on in the future, like Reed 1997.
They then discuss that drafting tools needs to account for how one set of information then have impact on other information. (Note again for me visual contracts seems very well suited to this) and exemplify with a narrow set of prices that are determined by certain combinations of time and events and the need to establish all set of conditions and outcomes that can occur and if neither are in conflict with any of the terms in the agreement. (Note again visual data!)
They then discuss the normative aspects and again a litterature reference I should check out as they have another approach that is not deontic logic only.
It seems a similar approach that what Pace & Schneider suggest in Article 7 (OPP-logic) but I should investigate further. In any case it highlights that many thinkers in the field believe a strictly deontic logic is not sufficient for complex contracts. The authors also discuss time aspects of contracts in a similar vein to in Article 8.
They then further investigate contract drafting into macro or micro level.
For drafting at micro level a symbolic language is proposed and a reference is to Article 1 (from 1957!) and I agree with the authors this seems a well suited method as a symbol language will allow precise meaning. And the authors even suggest some of these methods where using in legislation in the USA already.
The authors suggest on a Macro level the contract needs to be computer aided. I agree with this. My note: For example visual data of contracts where the computer and the blockchain is a tool for easy and efficient contract negotiations seems the way to go. In the future machine learning can probably infer more like for example intents or other less explicitly stated terms by learning patterns from large data of contracts but to get there we need first to AID humans and use computation as an aid for macro level contract negotiations.
The authors suggest you could start out with generic documents that can be evolved. I agree with this approach and many others are developing legal templates that can be extended. They also discuss how these documents can grow over time as new versions of templates are improved. (note: In line with my thinking on that a new contract field is going to be contracts over time. While before we could say redraft a contract every 5 years we can now do it in seconds or heck even milliseconds if automated far in the future. This again will allow us to much quicker learn from such contracts and particularly with machine learning.)
They propose a structure for such generic documents (note: templates)
Basically splitting between a layered approach of the scemantic meaning of contracts from abstract to finer details and then the information containing the relationship between sub components of the contract.
They also propose identifiers so that a document can be recreated by a fragment of the text together with a version number and identify of contract (Note: so a template with an address and version number)
They argue that you can then have building blocks (note: As in my own thinking that contracts can become like visual lego blocks at one abstraction layer) that can be layered into finer detail if needed.
Takeaway: Complex contracts and contracts over time can greatly benefit from computer aid. While expressive terms can be expressed in logical languages the more implicit information and the drafting of contracts on a macro level is best served with the computer aiding the user in the data and the user being the one in charge at our current understanding and tools available to represent contracts on computers.
1 Like
So, how do people intend to enforce force majeur clauses again?
Great question and one on my mind as well. I have not yet encountered any articles touching this in any detail but I am sure there are legal examples on this matter given how many digital goods there are. So it will be my thoughts since I have not read any on this yet: I believe this is best solved with a 3rd party validator but I am up for suggestions. Some clauses can be validated automatically for example lets say there is a volcano eruption with smoke causing all flights to a grinding halt then terms of immediate delivery by airplane of goods x cannot be done, this could either be verified by a neutral 3rd party (for example a legal firm on the block basically) or by aggregating data from several sources all confirming this to have been the case (weather data etc) this could be set in the contract in the draft period as in for example we agree that weather provider X weather data Y can trigger a force majeur clause Z. For me these contracts that require outside validations will become the more advance contracts and I want to start on the simple “invoice” type contracts or simple organizational contracts as I do not want to step on Marlow’s toes that only need the two parties to agree to make the contract form. Interestingly enough contracts over time would then be able to learn from failed contracts and add such force majeur clauses and could very quickly update terms and at zero to almost no cost or if involving a 3rd party as draftor much less cost as the visualization and data drive way the contracts are made will make the job much easier for the lawyers.
Validators is actually a pretty big topic in translating from the physical world to on the block for such contracts. Mobiles that most people have have probably some potential in this regard and I am sure we will see a lot of ingenuity in this field opening up even more advanced contracts. Likely in the beginning many of these things will be semi automated and with the power of crowd or many data sources to validate from physical world (as in the example of the agreed upon weather data as a neutral data point.)
The enforcement part itself is not that tricky if it is only financial information from a technical standpoint and from a pure legal standpoint I think it should be possible to adapt for different national laws on the fly with the CNL. If it is digital goods its somewhat more complicated (patents like say if someone keep using your picture even if you have had a void of the agreement, rights etc but I need to reread some of my law understanding of this to give a better answer) but I believe in many jurisdictions this can be programmed into pure blockchain contract movements instead of having to go to the physical world legal system. If it is not digital goods then you need to have your local nationally translated copy of the agreement (with the draft already specifying it applies to country X laws) and basically enforeced either by a 3rd party in private or by the national legal system. This is also why I believe a dualistic system is the optimal path for bitlaws and this is why I believe a CNL (controlled natural language semantic system) is the best path as it will allow instant translation of contracts into country X Y in a 2 party contract or country X Y Z in a 3 party contract etc. Just as EU laws are implemented in national laws you build with lego blocks contracts that are valid by national laws in the countries that are using the contracts - coexisting with national laws for anyone agreeing to use them. Of course this requires an identity layer as well on Cardano. But again for digital goods there could be easier ways, definitively for purely financial transactions. For actual physical goods you would probably be required to use the good old legal system. Another potential form of this is some middle men in the physical world that agree and guarantee for the flow of such goods in agreement with such contacts. Probably it will first be private and then nationalized as the benefits are more and more seen in all of society.
1 Like
Cardano community reads Computational Law Article 10
One Rule Extraction from Regulations is an article by Adam Wyner and Wim Peters
( http://wyner.info/research/Papers/WynerPetersJURIX2011.pdf ) where the authors discuss that rules have already been formalized with conditional and deontic logic but this has been a tedious process to do manually (with examples like RuleML) and proposes a natural language processing (NLP) technique to formalize national laws. (Note some form of processing of national laws with some manual checking seems optimal as with transnational systems like Cardano you will likely have a lot of legal changes per year across the world your serving, and would be a major benefit to any service provider on the Cardano network that provide contracts that are both in compliance with national laws so can be enforced and also automatically translated and in compliance between the national laws.) The authors suggest there has not been much success in machine learning based parser on this and another idea is to use a more linguistically approaches (Note: I agree with this, just look at all the stuff IBM Watson can do and it is also has linguistic parsing.)
The authors suggest currently available linguistic parsers are not well suited, among other reasons because they are not good at translating texts into deontic logic as for example the Standford Parser. They propose using an open source tool General Architecture for Text Engineering (GATE) that they use to:
(Note: Should be a good start with trying to parse some contracts to be used on Cardano for example governance based type of contracts.)
The authors then discuss use cases and how it is important to update and be in compliance with national laws and how this type of automated approach makes this process easier. They propose that a 3rd party can monitor and create legal rulebooks that can be computed upon and stored in for example XML format like LegalRuleML. (Note: This could be an interesting service provider on Cardano network that can benefit all type of contracts and any governance / legal framework would probably be wise to have built in importing rules for importing other’s formalized legal rules.)
To check if the tools and model they use work they selected a passage from the US Code of Federal Regulations, US Food and Drug Administration, Department of Health and Human Services
regulation for blood banks on testing requirements for communicable disease agents in
human blood, Title 21 part 610 section 40. This is a 1777 word document.
They then parsed it first with the Standford Parser to tag the text with correct linguistics characterization and as a plugin to GATE. However it failed to parse and the authors hypnotized several causes but also useful to know when dividing up the next they parsed it “successfully”. They had to manually check as well as it did not correctly identify all the text. The authors the propose that with the current state of the art knowledge it is better to make a semi automated process with manual checking during translation. They then use the GATE with Java Annotation Pattern Engine (JAPE) and rule based and again semi automated where a human oversees the process and steps in if the computer gets it wrong. Results where quite promicing
In general there was a high precision in identifying correctly but then to recall the connected words and rules it applied to was in some cases harder.
Key takeaway: Machine learning or Language parse based machine parsing systems can help in accelerating the formalization of legal text and perhaps also contracts but are still a way from being reliable. For the time being perhaps a better approach seems to be to use a CNL and translate rules from a known rule template.
1 Like
Cardano community reads Computational Law Article 11
(https://pdfs.semanticscholar.org/ebca/1f32ac887a975b60552920b9b1be619a87a2.pdf)
Modelling Contracts Using RuleML is a research paper by Guido Governatori and Antonino Rotolo.
Regarding ruleML the paper states: “RuleML is an XML based language for the representation of rules. It offers facilities to specify different types of rules from derivation rules to transformation rules to reaction rules.
Moreover it is capable of specifying queries and inferences in Web ontologies, mappings
between Web ontologies, and dynamic Web behaviours of workflows, services, and agent” further into the paper the authors expects ruleML to be used in distributed systems "It is expected that RuleML will be the declarative method to describe rules on the Web and distributed systems "
The paper focuses on making contracts machine readable by the application Deontic and Defeasible Logic and uses this for contract monitoring with ruleML. (From another paper on what defeasible logic is: Defeasible Logic supports rule-based reasoning where rules may be defeated by other rules that support a contrary conclusion. The concept of defeat lies at the heart of
defeasible reasoning. Where conflicts between rules arise, priorities can be used to resolve
these conflicts.)
A good example of defeasible logic use in the paper is ““every person has the capacity to perform legal acts
to the extent that the law does not provide otherwise” and where for example if you are a minor this is given priority to the original rule of “every person has the capacity (…)”
The authors acknowledge the ruleML is not without its limitations “It does not support explicit reasoning on deontic concepts and is unable to identify the behaviour of roles in the contract and contract violations.”.
They also discuss a form of logic needed for dealing with normative violations to make formal rules for such violations. “To reason on violations we have to represent contrary to-duties (CTDs) or reparational obligations.” (Note same concept that was more developed in the 2016 article 6 on OPP-Logic)
The article then describes how to use ruleML to represent rules and contracts and after the stage of machine language representation that one can process this mathematically into all outcomes then normalize this so any redundant rules are removed it can be put into the engine that monitors and runs the ruleML xml.
Takeaways: It seems possible to formalize the logical operations of contract rules by applying different logical approaches and this can then be represented by machine language. XML such as ruleML seems well suited for this and also for easily generating visual representation of contracts for ease of use for the average user.
Another area where Bitlaw would create new possibilities is in the field of DemTech. (See for example this https://seekingalpha.com/article/4231496-bitcoin-democracy-tech)
Imagine a world where we create better financial incentives to better the world. We have already taken baby steps in this for example in Emission Trading. (https://en.wikipedia.org/wiki/Emissions_trading)
But what if we combine this with Cardano based sidechain tokens that have governance features of bitlaw? We could create projects that decentralized governance and made it easy for every day people to create countless new types of projects and even new types of economies as in for example one with incentives for environmental sustainability.
(Draft 1. Picture by me.)
(1) Bitlaw offers templates that can be used to build contracts. Contracts can be individualized by a user or by a 3rd party that has fee in LAW tokens for example a ledger based law firm (LBLF). Bitlaw handles tracing of potential conflict between contracts using deontic law to computer assist the user or 3rd party in drafting contract. Drafting of contracts is handled through a visualization engine that is powered by a controlled natural language engine (CNL see article 5). Bitlaw could also learn potential pitfalls and conflicts in contracts through machine learning of large number of contracts or AI learning. Bitlaw handles all enforcments and monitoring that can be run through computations with IELE and Cardano SL. When needed to manually enforce through for example national police it can provide user with the contract portion needed to forward to authorities or also to forward it to a LBLF for enforcement or monitoring.
(2) Smart contracts through IELE and Cardano SL handles issuing of tokens through sidechain for example a Emission token (EMI) in a Demtech based project. A main contracts could be a Contract over Time (COT) that can learn from local contracts for example where a community energy provider creates a main contract for users to found local power projects implemented by said energy provider. The energy provider for example handles drafting globally for example with a LBLF (1) and locally with the users through the Bitlaw legal building blocks engine.
(3) Users create local contracts and transfer founds and distribute founds between users with Bitlaw (that uses IELE and Cardano SL). For example a local project #1 creates emission tokens that Community Energy Provider gets X if providing Y (verified in (1)). With a layered approach then implementation can be handled by a 3rd party for example Community Energy provider #1 while local users can have contract for distribution of founds and local voting on next steps of project etc.
What are some advantages of this compared with just making a contract?
- You gain 3rd party validation of contracts of more complex form than simple settlement contracts. (effectively paving the way for computational democracies and contracts over time that can learn)
- Barrier for creating such contracts are lower and ease of use is higher.
- You gain automation of many aspects of contracts from assisted drafting to in many cases automatic monitoring and enforcement.
- You gain lower barrier to legal assistants with 3rd party legal providers that works on legal templates effectively globalizing legal work same way internet globalized communication. These drafters gain ease of use through CNL and Machine learning.
- You allow for users to create own incentive systems that can spark new forms of economies instead of only growth based. (sustainability based or impact investor based etc.)
1 Like
I love this thread.
There is so much focus on how Cardano will revolutionize the world via the next generation financial stack, but so little on how the next generation legal stack will look like, while it will have at least as much if not more impact on our society, culture and environment.
Keep on writing your ideas, hopefully it will be a new roadmap item one day, maybe with a new DSL.
5 Likes
A long way to go as this is truly the baby steps. But maybe one day with persistence and community.
1 Like
1 Like
1 Like
Thanks. I am still interested in this topic for the longer run. I see quite clearly it is possible to create a legal DSL and that it could have major benefits for society.
1 Like
AMAZING POST… Im going to do a deep dive on this concept this week (and take notes on your post) thanks for sharing!
1 Like
Hey thank you for reading and reviving this post. I am still interested in the topic
This concept could prove to be the most consequential in all of blockchain… I have had many ideas on the topic bouncing in my head but didn’t know where to begin… This post is an incredible launchpad for what could become the first governance system that truly represents the mass of humanity’s virtues and morals in real time… I look forward to digging into the materials you provided… Please, keep up the great work!
1 Like
Dear @Eystein_Hansen ,
As a law student in the USA, I found this post and thread interesting! This idea would work great for elemental legal doctrines.
Your focus seems to be split between philosophical debates and legal doctrines. “IF/THEN” statements can be easily translated into computer language and programming, but things like “reasonableness” or “totality of the circumstances” or ‘freedom of speech’ or “substantiality” - may be harder to compute or illustrate.
I had a similar idea myself and found your post as a good launchpad to think more deeply about this field.
How is your research going lately? Pease post an update - I think you are on to something here.
1 Like