Should we replace judges with computers in contract cases?

The recent publication of another visionary book from Professor Richard Susskind has stimulated debate in the legal world on whether automation is going to put us all out of a job. One of the jobs that lawyers do is interpreting contracts.

IP Draughts remembers being challenged by scientist colleagues, over 30 years ago, on whether a computer would interpret contracts more accurately and consistently than human judges and lawyers. Instinctively, IP Draughts felt it would be too difficult to program a computer to address all of the issues that a judge needs to consider when deciding a contract case.

The scientists were disappointed that it wasn’t possible to have certainty on how words used in a contract would be interpreted. This desire for certainty seemed to assume that the task of interpretation was simply one of applying linguistic analysis to the words used.

Nowadays, IP Draughts is becoming increasingly frustrated with how courts sometimes seem to ignore the linguistic meaning of contract terms and instead interpret the contract in a forced way, in order to do justice in the case. He put this point to a Court of Appeal judge recently at a UCL dinner. The judge commented that judges were always going to try to do justice in the case, as it was their job to do so, and if IP Draughts didn’t like that he should provide for arbitration in his agreements. Later, IP Draughts reflected that he wasn’t sure that arbitrators as a class would be any better at disciplined linguistic interpretation than judges.

Meanwhile, Ken Adams has recently raised the question of how often a case turns on the interpretation of the words used. He hopes to arrange for some empirical research to be done on this topic.

These varied thoughts are prompted by reading the judgment of the Court of Appeal in the case of Mortgage Express v Countrywide Surveyors Ltd [2015] EWCA Civ 1110, which was issued last week. The case turns on a small point of contractual interpretation.

The case concerned claims by a mortgage company (lender) against a surveyor in respect of the latter’s valuation of properties. The parties had entered into a standstill agreement, under which they had agreed that the clock would be stopped in relation to time limits for bringing an action. The present action was over the interpretation of the standstill clause. The key part of the clause read as follows:

2.1. For all purposes of any defence or argument based on limitation, time bar, laches, delay or related issue in connection with the Dispute (a ‘Limitation Defence’), time will be suspended from the date of this Agreement until 30 days after the service by any Party of a notice which is compliant with Clause 3 below stating that the running of time is to recommence (the ‘Standstill Period’).

A recital to the agreement defined Dispute as follows:

4. In this Agreement, ‘Dispute’ means any claim or claims directly or indirectly arising out of or in any way connected with the matters referred to in paragraphs 1, 2 and 3 above.

Paragraphs 1 to 3 summarised that the defendant had been hired to conduct valuations and that the claimant had alleged (in 2010) that the valuations had been performed negligently.

A dispute arose as to whether a later claim, in 2013, was covered by the standstill agreement or was time-barred. The later claim, which was concerned with the same valuations, alleged “fraudulent misrepresentation and/or deceit and/or breach of contract and breach of duty of care”. The defendant argued that a claim in deceit was not contemplated by the earlier settlement agreement and was time-barred.

The recitals to the standstill agreement mentioned negligence but not deceit.The question of interpretation was therefore whether an allegation of deceit was , in the words of the fourth recital, “directly or indirectly arising out of or in any way connected with the matters referred to in [the recitals]”.

At first instance in the Technology and Construction Court, HH Judge Raeside QC agreed with the defendant. He commented:

I do not see how a claim in deceit can either directly or indirectly arise out of (1) and (2) and (3) of [the recitals]. The deceit claim is a case of systematic, opportunistic and deceitful overpricing from the outset and over the full period of time by these surveyors of each and every one of the 50 properties …

HH Judge Raeside QC (not to be confused with HH Judge Raeside)

It seems to IP Draughts that this interpretation chooses to ignore the broad wording of the fourth recital, quoted above. The judge’s reasoning could be viewed as saying that deceit is much worse conduct than negligence, and the language used by the parties should be subject to an implied term that limits the words “directly or indirectly connected” to claims based on the same types of causes of action. It seems to IP Draughts unlikely that any computer would be programmed to come up with an interpretation of this kind.

The Court of Appeal overruled Judge Raeside. In the words of Lord Justice Simon, who gave the lead judgment:

…in my judgment the wide definition of ‘Dispute’ plainly extends beyond what had been alleged by the First Claimant at that stage. It would, for example, plainly cover a claim for breach of fiduciary duty based on the payment of a bribe. The difficulty with Mr Lawrence’s reading down of the paragraph 4 is that it could have been achieved by a much shorter paragraph 4: ‘In this Agreement, ‘Dispute’ means any claim or claims directly or indirectly referred to in paragraphs 1, 2 and 3.’ His answer, that the Claimant might have wanted flexibility in relation to particular valuations does not, in my view, give sufficient weight to the words ‘in any way connected with.’

In my judgment the proper construction of the Standstill Agreement is that if the claims arise ‘indirectly’ from the matters referred to in paragraphs 1-2 of the Background Preamble, or if they were in some way connected to those matters, they fall within the suspension provisions. The claims based on dishonesty fall within this very broad category of claims since they were at least in some way connected with the factual matters set out in paragraphs 1 and 2 of the Background Preamble and with the specific allegations described in paragraph 3.

Good. But this small case illustrates how much scope there is for a judge to come up with his or her own interpretation of the words used in a contract. Does this mean we should apply a simpler set of criteria to interpreting the words, based purely on linguistic analysis, which a computer might be able to replicate? IP Draughts is still undecided.