A scientific study has found that contracts are hard to understand not because the ideas inside them are too complex for people to grasp, but that lawyers write in such a way as to be incomprehensible to most of the world.
The findings will no doubt encourage the work of legal design experts, while the study will also have relevance to NLP companies trying to make contract analysis easier, as well as those interested in legal standards and ‘super-structured’ contracts.
The study, led by Eric Martínez at the Brain and Cognitive Sciences group of the renowned Massachusetts Institute of Technology, studied a corpus of contracts totalling around 10 million words, and found that: ‘Contracts contain startlingly high proportions of certain difficult-to-process features, including low-frequency jargon, centre-embedded clauses (leading to long-distance syntactic dependencies), passive voice structures, and non-standard capitalisation relative to nine other baseline genres of written and spoken English.’
Or, to put it simply, lawyers break every rule in the book when it comes to clear and direct writing. Some of the reasons are obvious, e.g. lots of obscure jargon. But others are really interesting as they show that legal design needs to tackle syntax (how words are positioned to create a sentence) as much as anything else.
For example, one key problem is the use of clauses that are dependent upon other clauses for you to understand what’s going on, (i.e. a bit like the style you will find when reading some Victorian novels). Also the use of the passive voice – a total no-no in journalism and also in thriller writing, (i.e. ‘Mary opens the door’ which is active, not ‘The door is opened by Mary’). The passive voice slows things down and also makes it hard for the reader to follow the main subject of the text.
The study also stressed that readers can understand the concepts and overall ideas used in a contract, e.g. A owes B money, C will pay a ‘fine’ to D if something is late, E will be responsible for looking after F area of concern etc. These are in fact all basic human concepts and people understand them. What happens is that when this is drafted by a lawyer it becomes unintelligible to anyone except another lawyer for the reasons mentioned above, i.e. a mix of jargon and an impenetrable writing style, turns what should be understandable into something that switches the brain off in those not inculcated in the arts of legalese.
The research has come to light after winning an Ig Nobel award, a tongue in cheek prize for unusual or obscure scientific efforts.
Of course, many will say that they know this already. But, if that is so, then why do contracts still feel like they are written with the express purpose of being hard to understand? Is this some unconscious guild protection instinct, i.e. if we write in this way we protect the guild as no-one but us can understand what we do?
Or, is it just bad habits that started hundreds of years ago and still have not been addressed…?
In either case, this makes it hard for the rest of a company, outside of the inhouse legal team, to understand these key documents, which are essentially the DNA of the business.
It also means that NLP software really has to strain to make sense of legalese. Many vanilla language models seem to do well on everyday texts, but don’t do well on legal docs without plenty of additional specialised training – and that in itself tells us something important.
So, if contract writing could be improved, just think of the benefits it would bring, to legal tech and the wider economy….! Is it time to bring in a back to school programme….?
Here’s a link to the study.