TOWARD A NEW LANGUAGE OF LEGAL DRAFTING.

• Author: Roach, Matthew

Contents

1. The Problem

   1.1. what got me thinking

   1.2. How lawyers draft and publish contracts

   1.3. Why this is a problem

2. Solution

   2.1. Thinking about the roles that contracts perform

   2.2. More functional contracts

3. What This Could Look Like

   3.1. Key Elements

   3.2. Authoring in a legal markup language

   3.3. Sharing content like a coder

   3.4. What it could look like in practice: construction contracts

4. Possible Objections

   4.1. Lawyers can't or won't draft like this

   4.2. This won't be effective without industry-wide standards

   4.3. This can be done as well or better within traditional word processors

   4.4. The structure is vulnerable to changes in technology

5. How This Compares to What Others are Doing

6. Conclusion

   List of Figures

1. Example clause

2. Example clause with markup

3. How a software interface can refer the author to other elements in their document, as they type

4. Example of a change to a HTML document tracked by GitHub

5. Screenshot of ContractExpress Author (BusinessIntegrity 2015)

   List of Tables

1. Description of tags used in Figure

2. Examples of legal XML development

3. Examples of companies providing contract automation services

4. Examples of companies providing legal data analytics

1. The Problem

   1.1. What got me thinking

   A common experience of being a lawyer that you don't think much about process improvement or product design. The key focus for many lawyers is meeting client needs as quickly as possible and billable hour targets. Having been a transactional lawyer for several years, I had never thought of drafting contracts in anything other than Microsoft Word.

   When I started my LLM I met math and engineering students, who were involved in various forms of data analytics, machine learning and natural language processing. They showed me their projects and the software tools they were using. I realized that in other disciplines, people are adept at switching between the languages of math, coding or natural language, often within a single document, in order to use the tool best adapted to the task at hand.

   Taking classes in design, technology and law, I began to think about the potential for changing how we generate and access legal content. I began to reflect on how we access content in various forms through technology, and how far the design and accessibility of legal content lags behind what we now take for granted everywhere else. This paper explores the thought that there is an enormous potential functionality that can be added to legal content if lawyers make modest efforts to add machine readable structure to their drafting. Lawyers would enjoy learning new skills, and clients and lawyers alike would be excited to discover how the way they produce and access legal content could be transformed.

   This paper discusses what authoring in a markup language might look like, some of the advantages that this could have, and some of the barriers to implementation. A related question is what it would take to shift lawyer behavior to this style of writing, and what transitional steps might be appropriate. This could be the subject of further work.

   1.2. How lawyers draft and publish contracts

   Lawyers draft documents in word processors that focus on formatting and final appearance, usually Microsoft Word. Their documents are almost universally accessible and editable by the lawyer's clients, the other side and the courts. Following initial preparation by a lawyer, a draft contract may be emailed back and forth many times, with the parties making and tracking various changes.

   once the parties agree the terms, a junior associate tidies up the formatting of the document, prints it out and walks around town getting it signed. If you're super modern, you might do electronic signatures. Then the associate makes a pdf, emails it around and everyone uses that or the final word document for ever after as the record of the deal struck.

   1.3. Why this is a problem

   The output of legal drafting as it is done now and has been done in the past is unstructured natural language, poorly adapted for computational use and analysis. This is a problem because:

   * Outside law, clients and lawyers are used to accessing and editing content in much more user friendly formats. (1) They use web based platforms that look great, allow easy navigation and transformation of content presentation according to the user's needs. (2) They can do this because the content has embedded structures readable by computers. (3) * Sharing and editing content using traditional word processors is cumbersome compared to tools used to create, edit and share digital content outside law. (4) GitHub and similar tools used by coders offer greater flexibility and functionality. (5) There are significant productivity and user experience costs to keeping outdated tools just because they have become familiar and something of an industry standard in law. (6) * Contracts are poorly integrated into other business and data management systems, and companies don't know enough about the contracts they have entered into. (7) As Nick West of Axiom has noted "very few general counsel can tell you the number of contracts their company is party to, let alone understand the totality of their obligations, the interactions between them or their organizational risk implications." (8) * Lawyers aren't managing their knowledge and experience effectively. (9) A significant part of the practice of law is drawing on experience and knowledge gained from previous transactions and documents. (10) Legal drafting often involves a lawyer taking a moderately well-structured precedent and customizing it into an unstructured or "flattened" form. (11) It is hard to force the output back into the knowledge management system. (12) In most cases the lawyer doesn't bother, which affects both the quality and efficiency of future work. (13) * While machine learning and natural language processing techniques are improving, the ability to undertake computational analysis of legal documents is significantly complicated by their lack of structure. (14) If the lawyer is conscientious in how they use their firm styles it may be possible to parse a basic structure out of legal document and identify what are section headings, defined terms, legislation, case names, etc. (15) But this is hard work and unreliable. (16) It would be much better to start with something structured. (17) * An outdated approach to content generation and publishing is part of the legal industry's broader vulnerability to changes in technology and new legal business models. (18) Clients are fed up with the traditional law firm. (19) Lawyers cost too much and aren't productive enough. (20) Lawyers haven't looked outside law to see what's happening in the world. (21) Lawyers need to learn some new tricks, and start to catch up with everyone else. (22) 2. A Solution

   2.1. Thinking about the roles that contracts perform

   Lawyers author and publish contracts as if their only purpose is to be a permanent record of the parties' bargain, to be kept in safe

   storage and dug out and presented to the court if absolutely necessary. (23)

   But many contracts are living documents that go through a period of evolution as they are drafted and negotiated, and then use and reuse after they are signed. (24)

   Contracts perform a number of roles, and have a variety of data and business knowledge embedded within them. (25) The roles of contracts include:

   * A platform for negotiation and collaboration. Many commercial aspects of deals are conceived or refined through the formal process of reducing the parties' deal to contractual provisions. (26) * A store of information about the deal. This may appear in a contract in various forms, but may include classes of information such as monetary amounts, dates, references to external events, etc. (27) * A store of knowledge about the parties' businesses. A variety of business data may be embedded within a contract, from simple things like address and contact details, to more detailed commercial information such as insurance policies, business procedures (e.g. ways of making payments), etc. (28) * A store of knowledge about the law. Much of a lawyer's value comes from the experience of previous transactions. The content of contracts are a mineable resource for future deals. (29) * A reference for the court, should a dispute arise. (30) 2.2.More functional contracts

   Having regard to these different roles of contracts, it is worth considering whether the current approach to authoring and publishing contracts takes full advantage of available technologies to maximize their value and usefulness. (31) Imagine content authored by lawyers being spun out seamlessly to clients, supervising partners and other parties in a form adapted to their needs and use. Only need to see one clause in a 200 page contract? See the clause, its history, edit it and send it around without wading through the rest. (32) Your partner can approve your clause amendments on their phone in a cafe just as easily as they can in the office. (33) We could see genuine collaboration between teams of lawyers and their clients in legal content creation. (34) Have suggested definitions and clauses appear as you draft. (35) Have the software check that you haven't defined a term then not used it, or vice versa. (36) Hover over terms and see their meaning, change histories, relationship to other parts of the document, or other docu-ments. (37) Your document is effortlessly available as a precedent for later transactions. (38)

   This could be achieved within a software environment based on lawyers authoring in a document markup language like XML, HTML or LATEX, specifically adapted to law. (39) They should also take tools from the coder's toolbox like GitHub to help them share, edit and record change histories of documents. (40)

   This is not hard. (41) Lawyers would need to produce content in a way unfamiliar to them now. (42) But other professionals have...