Keywords

1 Introduction

Contract visualisation has become a vivid area of innovation engaging lawyers, researchers, designers, and practitioners alike. The different motivations for so-called “contract simplification” gravitate around the aim to enhance understandability and transparency of contracts. As elaborated in previous research [1, 2] we believe, that there exists oftentimes during the process of contracting a gap between the legal representation of an agreement and the goals and intentions of the negotiators.

Helena Haapio [3], an academic and lawyer herself spearheading the proactive law initiative states that “Proactive lawyering is not about applying legal rules to facts that happened in the past, but about applying sound legal practices to create future facts and to plan a future course of conduct”, ideally preventing litigation. An important aspect of preventive law is to enhance the understanding of a legal setting before and during the negotiation, before signing a contract (therefore understanding its implications) and after in case of violations of the agreement.

With this paper, we aim to develop a general visual approach supporting the various stages relevant to contracting. We will start with a short overview of current practices in the area of “contract simplification”. These include i.e. abstract visual languages and icons [4] which may need to be learnt, narratives like “comics contracts” [5] which may underrepresent the intended complexities of an agreement, but could work well with illiterate legal partners (e.g. in developing countries). The approach to illustrate specific situations relevant to the contractual content may prove inflexible (and therefore potentially unsalable) if they need to be created for each context from scratch. Inspired by the existing approaches to visual contract simplification, we aim to develop a visual language which combines abstract and specific elements.

We develop three models of representing legal contracts. We introduce the concept of “possibility space” (model one and two) to display counterfactual “states of the worlds” [7] which contracts aim to organize. For the last model, we assume that contracts can be reduced to core modal categories, such as obligations, prohibitions, permissions and the concept of options (“no obligation”) [6]. These rather high level abstractions are the core entities relating to the parties involved.

2 Current Practices of Contract Simplification

Under the concept of “proactive law” [3] legal academics, practitioners, and policymakers have identified and described increasingly the deficiencies in current legal practice and the opportunities for innovation. One main concern is the lack of transparency of legal documents and the focus on technicalities rather than on the actual deal-design of an agreement.

Under the concept of “contract simplification” a range of ideas and approaches have surfaced ranging from linguistic simplifications (natural language), reductions to code (if law equals somehow computer code) and approaches for simplifications via visualisations and graphical means.

2.1 From Language to Code

There are different kinds of linguistic representations for contracts [8], see Fig. 1.

Fig. 1.
figure 1

Natural language progression

In Fig. 1, Martin describes a natural progression of language goes from prose language in contracts to Standard English (formal language) contracts to auditable language meaning language that could be analysed by computers. From there we progress to the formation of contracts using high-level machine readable languages. In the next step, contracts are coded by machines. To achieve full automation of the legal process, regulatory and enforceability issues across legislations need to be resolved.

Due to the blockchain technology and the Ethereum project [9] built upon that technology, the time for the real automation in contracts might arrive soon. The vision behind the Ethereum platform is to facilitate building software on the “blockchain”. The blockchain technology is a chain” that is a distributed ledger in the form of databases, whereby the distribution prevents fraudulent changes within this digital record. Blockchains can be public, but don’t have to be. For the vision of “smart contracts” as proposed by Nick Szabo [10] all information is public and therefore transparent for all members of society. Szabo (and others, such as Meng Wong [11] from Legalese) view contracts as a set of “if-then rules”, which therefore are prone to be represented via code (computer code and programming languages) and not natural language or legal lingo.

Through this technology contracts can become self-executing – similar to all software: once the premise of a rule if fulfilled, the consequence can be triggered automatically. The vision of smart contracts as self-automated contracts for the better of society preventing human errors and flaws [10] have been widely discussed and serious concerns have been raised. One of these is of course transparency and control of such software programs.

Coming back to the role of contract simplification and the role visualisation may play: smart contracts – provided they are meant to serve humans will need to be controlled and understood in their complexities. We believe there is huge potential even for less radical approaches to self-executing software in the form of smart contracts.

2.2 From Language to Visualization

Images and visualizations [12] can convey complex problems in a simpler way and increase understandability [1]. Hence, it seems beneficial that if we could find out of way to visualize contracts, it would benefit all the parties as they are now easily comprehended with all their consequences. This also brings us back to the idea of “proactive law” [3], instead of focusing on the technicalities of contracts and minimising damages in situation of violation, better understanding of the contracts arguably helps the design of deals and mitigates the possibility of violation in the first place.

Visualization of contracts can have been discussed in different forms. In “Next Generation Deal Design”, Haapio, Plewe and De Rooy [5] discuss different forms of visualizations. Namely, visualization as contracts, visualization for contracts, visualization in contracts and visualization about contracts. They all serve different purposes and some might be arguably more practical in real life. Furthermore, under visualization as contracts, the idea of using comic illustrations as contracts are brought about. However, we believe that the practicality of visualizations and their creation is an important point and therefore favour approaches where the visualisation is somehow automated [3] as well as simple to understand.

3 Existing Approaches

Recent research explores various categories of visualizing contracts. Haapio, Plewe, De Rooy [5] introduce at least four ways: visualization as contracts, visualization for contracts, visualization in contracts and visualization about contracts.

Mahler points out that there is no single icon that can be understood universally [6]. Hence, educational cost of helping people to understand all the icons tend to be high. For example, the concept of authority is common in contracts, but there is not a universal icon that every potential user could understand. We would think of using the icon of a crown to symbolize “authority”, but not only the crown icon is different throughout the world, not everyone understands what a crown is.

De Rooy proposes to convey the contractual dialogue between parties via the combination of text and visuals in the form of contracts. [5] He uses comic drawings to represent possible scenarios in the case of an employee agreement. Most possible job requirements and consequences are captured using those drawings, however, the comic contracts have a challenge: hard to automate and therefore commercially hard to scale. Hence, it may have limited use cases, unless there is a way to translate text into comic-images in an automated way. Considering the current state of AI this seems hard to realize in the near-term future. However, comic contract once generated can have tremendous social impact and help to empower illiterate segments of the population.

Most of the existing proposed solutions rely on categorizations and conventional symbols. Potential limitations of such strategies are that Icon families are oftentimes not systematic oftentimes and can’t be easily extended and are therefore potentially inflexible. Another approach is introduced by Passera et al. [4] where she enriches contracts with automated visualization mostly representing quantitative data, for example, she illustrates in an interactive chart a payment clause with its penalties in case of the violation of the agreement.

4 Models of Visualization

For our first two visualisations we introduce the concept of “possibility space” By possibility space we mean the “possible worlds” which if-then rules are describing. This understanding is inspired by David Lewis Possible World semantics [13]. We use his modal semantics as a kind of metaphor to craft a visualisations, which can express contracts the way they organize the future states of the world.

One obvious approach is to organize the possibilities of the future along in a chronological fashion with a timeline and some branching universes. Visualising Lewis Possible worlds would probably look like this: he aims to define the semantics of modal logics via various possible worlds with so-called “accessibility relations” between them. So, for example, if in a world, A x is possible, then from world A can be accessed a world B in which x is true. From world A is also accessible a world C, in which x is not true. And – depending what degree of granularity one’s underlying logic allows for, any degree between true, no true- such as not known, the opposite is not known etc. We will use this kind of “accessibility- relation” between “Zustaende der Welt” (“states of the world”) [7] to describe the various scenarios in a branching time universe a contract aims to cover. However, this approach at least in theory suffers from the combinatorial explosion and therefore may only be useful when small possibility spaces are sufficient to show.

Since contracts have always an element of time to them, we will experiment with various other ways to capture time. Interactive media and its origin in the concept of “hypertexts” [14] supports this kind of thinking naturally, we believe. However, this approach may provide a false sense of security and may not capture the eventualities of the future as efficient as language can. We may be reminded through these sketches, that language as means of simplification (mostly through the power of abstraction) may be considered the most expressive and efficient medium – a reason, why contracts are probably till today conveyed in text.

Nevertheless, we will also explore a hybrid approach of combining text and spatial position of text (concept clustering) to reduce complexities. This approach will be based on representing the parties and their obligations, permissions and prohibitions and in that sense at least make the contract and the actionable components clear to all parties. We may conclude through our research that visualising the varieties of possibilities is perhaps not necessary and efficient (i.e. aesthetically elegant).

4.1 Model 1: Road Progression Model

This models puts emphasis on a timeline, analogically model the progress of a contract to the progress on the timeline. It conveys how one decision may have far reaching consequences. Three key component is represented in the road progression model, they are the representation of concepts, the key milestones as well as the end of the contract.

At the beginning of modelling, we define the icons used for modelling contracts, then we represent the contract with those icons, lastly, we explore what may cause the termination of the contracts. Here we use tenancy agreement to model the contract.

As shown in Fig. 2 a definition section where all the icons are defined upfront to avoid confusion in the future. Today’s contracts are complex and their meaning is not always clear to those who are impacted. What is presented above is a way to make contracts more engaging with icons replacing words.

Fig. 2.
figure 2

Road progression model interface 1

As shown above in Figs. 3 and 4, this model focuses on the importance of timeline, every contract has its life time. In the case of a public housing tenancy contract, 5 year can be a critical time as it is the time when the tenant can start renting out the house. This condition is represented through the middle block. If tenants choose to rent it out, the contract continues, else a new contract is required.

Fig. 3.
figure 3

Road progression model interface 2

Fig. 4.
figure 4

Road progression model interface 3

Furthermore, by making the interface interactive shown in the Fig. 4 the user may be able to click on a change of event to know the consequences of his actions. For example, if the tenant illegally sublets the house, this would lead to a premature termination of the contract and certain penalties may apply.

4.2 Model 2: Spanning Tree Model

This model maps out all the possibilities of the consequences that are caused by users’ actions. To illustrate the use of the model, our example is an insurance contract (Fig. 5):

Fig. 5.
figure 5

Spanning tree model interface

This is an example mapping out all the possibilities that may occur in the course of contracts. It leads from general terms to the more specific terms, covering all aspects both parties need to know about the contract. However, if one takes a look at the exclusion terms one may realize that in many possible ways, the contract would be void and the insured person would not get paid. If we start tracing the conditions to different terms, we would run into an explosion of possibilities branching out from one scenario to another.

Since insurance policies are usually based on a timeline, multiple scenarios might happen along the timeline and events happened before hand might affect the events happened after. In this particular example, the insurance policy states that it only insures up to four persons from the same family, hence, in the case of political evacuation, it means only four of the family members’ evacuation fees will be covered. On the visual representation, this requires more layers. Hence, more events would lead to a combinatorial explosion which is counter intuitive to the purpose of visualization as its aim is to simplify the contract. Nevertheless, this model maps out the possibility space and helps to navigate the contracts which is valuable for the understanding of the contracts.

4.3 Model 3: Modalities Model

Inspired by Lewis’s [13] Possible World Semantics to model modal logics and Mahler [6] work on visualising modalities, we will now introduce a visualisation based on high-level abstractions such as the four modalities “permission”, “prohibition”, “obligation” as well as “option”. We believe that by visualising abstractions such as modalities, we can avoid some of the drawbacks of other visualisations, which oftentimes tend to be too specific and therefore may not lead towards a universal visual language for contracts.

Similar to our previously proposed negotiation platform, in this concept, users first list the participating parties and a kind of market or negotiation space [15] between them. For each party, there is a “corridor of negotiation” [16] allowing for actions along the various modalities.

All terms in the contract can be categorised in the four categories above. Permission is the action of officially allowing someone to do a thing; consent or authorization. Prohibition is the action of forbidding something. Obligation is an act or course of action to which a person is morally or legally bound; a duty or commitment and option is the defined as no obligation for example the U turn sign, you are given the option to U turn. Top and bottom of the model represents both sides of the party, in this case, the landlord and the tenant. Each dot/item represents one action that a tenant is allowed, prohibited or obligated to do. When the user click on one dot, a floating sticker would appear to show the specific term that dot represents (Figs. 6 and 7).

Fig. 6.
figure 6

Modalities model interface 1

Fig. 7.
figure 7

Modalities model interface 2

In the example of a tenancy agreement - for each party the various items are defined and could – provided an interactive interface – be unfolded, looked upon, discussed, and agreed upon etc. by each party. In other worlds, this visualisation would become a strategic interface (Plewe) to facilitate also the negotiation process [15].

5 Comparison

We would evaluation the usefulness of these three models based on the criteria we have identified:

  1. 1.

    Supporting a visual language for contracting - reusable and universally applicable organizational principles or building blocks (e.g. diagram parts) usually describe best practice structures or concepts of specific domains which can be adapted and used in various contexts.

  2. 2.

    Capture the chronological nature of the possibility space – if necessary

  3. 3.

    Increasing understandability – before contracting and negotiation, during contract drafting and after the contract is signed, i.e. understand the implications within the possibility space

  4. 4.

    Facilitating users in different roles (e.g. managers, laypersons, software engineers or lawyers) often use a law or contract for different purposes. It is important for a legislator, contract designer or law-interpreting person to know the rationale behind a certain norm, contract or law.

The road progression model qualifies for the first three factors. Specific maps would be generated for specific group of people. In the case for Tenancy agreement, there would be two maps generated. One for tenant, one for landlord. As the landlord is the one with power, it is understandable that in the contract constructed, there would be fewer disadvantaged terms for him. However, in Singapore, the government does have terms that protects the rights of the tenant for example and does not allow short term renting. This should be considered in the landlord map. It is important that the model is modularised so that its interface can be used by code. Furthermore, this model is excellent at showing the change of event in time, this is modelled by the concept of ‘road’ where violations of certain terms would lead to different consequences. However, one disadvantage of this model is the categorization which cannot be standardised. It changes with the focus of the contracts which makes the automation of this model difficult.

Like the model above, the spanning trees model is good at the first three points. It is more intuitive than the Road Progression model as it aims to include all possible scenarios in its representation. However, actions that could be taken in sequence and this would lead to a combinatorial explosion that implies too many possibilities and makes it practically infeasible to visualize the whole possibility space.

The road progression model and the spanning tree model are both visualizing modalities implicitly. However, a major disadvantage is that they are not easily automated. There are hardly a standardised ways to exhaust all the possible concepts that could appear in one contract. This implies, all possible cases would need to be represented in the software before determining their interconnections. In the last model, modalities are being expressed explicitly, making it possible for the system to exhaust all terms of contracts.

The modalities model has a great advantage that is its clarity of representation. For both model above, they are categorised based on the terms that are used in the contract. They change when the focus of contract changes. But with deontic modalities, all the terms under this model are categorised by the nature of the action. This simplifies and consolidates the terms in a contract and makes the interface more logical. It would work well with code too. Its Drawback is, that users do have to learn the abstractions used as explained in Mahler’s paper [6] – but hopefully the modalities are intuitively to grasp.

6 Conclusion

On a philosophical note, all forms of representations are forms of reductions of complexity and suffer to some degree from inadequacies. This remains a challenge for any documentation of an agreement, be it as code, text or a combination of those aiming to anticipate future events.

The emergence of new legal technologies provide opportunities and also demand new and interactive solutions to facilitate and capture agreements between humans. Visualisations will become increasingly important to support not only human-human interactions but also computer-human interactions. Any constellation is imaginable, e.g. lawyers among each other, lawyers and deal-makers, laymen negotiating, peer to peer online contracting and any users of legal software.

Generic modality based interfaces as the here proposed model 3 could be promising candidates to support technology based contracting. It seems desirable, if interfaces cater independently from the domain to a variety of contracts without introducing specific visual elements (e.g. illustrations), icons (leading to rather inflexible visual language which might be difficult to extend/adapt ad hoc) or particular conventions (which may have to be “learnt”).

In the context of the blockchain enabled platform Ethereum modalities based interfaces such a model 3 could be implemented. When contracts can be read and analysed automatically, and the particular modalities, such as obligations, permissions and prohibitions can be identified automatically – then this approach could enable the automation of the visualizations within such a modal visual framework. We will explore such frameworks in further research.