I will keep my comments on what has been said during this rewarding Colloquium within a small compass for two reasons. The first is administrative. The proceedings must run to time. Secondly, an unexpected circumstance restricts my freedom of expression. At the time when I was honoured to be asked to participate in this Colloquium I could not know that I would become involved as an appellate judge in the large and contentious dispute concerning the Channel Tunnel, which as previous speakers have made clear, has given rise to issues falling precisely within the scope of our discussions. Naturally, I can say nothing which could reflect on anything said or to be said by the English court in respect of these matters. Within these constraints there are however some observations which I can properly make on this important subject.

I say that the subject is important for two reasons. The first is in a sense symbolical. For too long during recent decades there was a marked difference of theoretical opinion about the relationship between arbitration and the courts. This difference, sometimes sharp enough to resemble a confrontation, has a history which would merit a doctoral thesis and perhaps a full scale historical work. There is no time for this in ten minutes, and in any case the dispute itself has been the opposite of rewarding, since it has exhausted a great deal of time and effort which would have been better spent in the practical task of making arbitration a more effective servant of the international trading community. Happily, matters have greatly changed during the past few years. The tremendous scale and intensity of international disputes leave no room for the doctrinaire. What is urgently needed is that arbitrations are made to work and seen to work. Under these pressures arbitration has been brought back into contact with its roots, not as an abstract discipline like pure mathematics, but as a structure whose only raison d'être is to make international commerce more effective.

The choice of a subject, and the attraction today of a large audience for a subject, which addresses the practical problems of the interface between the courts and the process of arbitration is to my mind a most welcome signal that the combination of a down to earth and intellectually magnanimous approach to our chosen topic is now becoming recognised as the spirit in which the study of arbitration should be embodied.

The second reason why the subject discussed today is important, and why the fact that it is being discussed today is also important, is that both the framework of the Colloquium and the discussions which have taken place within it recognise that the arbitral process cannot remain effective without a partnership between that process and the courts. The old and sterile confrontation between the "minimalists" and the "maximalists" regarding the part to be played by the domestic courts has now given way to a recognition that the courts must recognise the essential role of arbitration in international commerce, and give it the maximum permissible support; and a converse recognition that arbitration cannot flourish without that support.

All this being said however there remains an inevitable tension between the duty of all concerned to honour the agreement of the parties to have their disputes resolved by a voluntary, private and autonomous mechanism, the frank recognition that there will be instances where through ill will or bad luck the agreement to arbitrate breaks down. Here the only alternative to throwing up one's hands in dismay or resorting to futile exhortation is to call in the court to make the parties comply by compulsion with an agreement which was intended to be spontaneous. This tension is at its minimum at the opposite ends of the process. Nobody could sensibly deny, more than thirty years after the New York Convention, that the autonomous arbitral process may need the help of the court at the outset, if a claimant starts legal proceedings in breach of his agreement to arbitrate or if the [Page119:] arrangements for constituting the tribunal break down; or at the other end if the losing party fails to honour the award. It is at its maximum where the jurisdictions of the arbitrators and the domestic court may overlap.

It is when one comes to the middle ground between the extremities that one finds a potential for a conflict between the roles of the arbitral tribunal and the domestic court. May I very briefly indicate what some of these tensions may be.

In the first place there may be a tension about the points at which one institution yields a dominant role to the other. Ideally, the handling of arbitrable disputes should resemble a relay-race. In the initial stages, before the arbitrators are seized of the dispute, the baton is in the grasp of the court; for at that stage there is no other organisation which could take steps to prevent the arbitration agreement from being ineffectual. When the arbitrators take charge they take over the baton and retain it until they have made an award. At this point, having no longer a function to fulfil, the arbitrators hand back the baton so that the court can in case of need lend its coercive powers to the enforcement of the award. But in real life the position is not so clear-cut. Very few commentators would now assert that the legitimate functions of the court entirely cease when the arbitrators receive the file, and conversely very few would doubt that there is a point at which the court takes on a purely subordinate role. But when does this happen? And what is the position at the further end of the process? Does the court retake the baton only if and when invited to enforce the award, or does it have functions to be exercised at an earlier stage, if something has gone wrong with the arbitration, by setting aside the award or intervening in some other way? That a tension does exist was clearly demonstrated by the debate about the inclusion of Article 34 in the UNCITRAL Model Law.

Other tensions arise even at a time when the arbitrators are firmly in the saddle. For instance, if there are steps which could usefully be taken to make the arbitrations more effective but the arbitrators have no power to order them, is it a proper function of the court (in its subordinate role) to step in and fill the gap; or is this an illegitimate usurpation of the right of the parties to choose an extra-judicial method of dispute resolution. The same question may be asked where the arbitrators do have the power to order the desirable steps, but lack the means (which the court does possess) to force the party to comply. If one starts with what I believe is now the most widely-accepted view, that the courts have a part to play, but that they should voluntarily abstain from all but the minimum necessary interference, one must ask "Which is preferable, to lend the coercive powers of the court to reinforce the authority of the arbitrators or to keep out of the arbitral arena in order to maintain the integrity of an autonomous voluntary process?"

This tension becomes particularly acute when the assistance of the court is legitimate in principle, but cannot be granted or withheld without requiring the court to enter into the substantive merits of the dispute. Conservatory seizure provides a good example. Even the most ardent minimalist would hardly deny that such a measure can properly be granted by the court in support of an arbitration. Yet in many jurisdictions the court will not grant such relief without a preliminary assessment to satisfy itself that the claim has real merit. Should the court be performing such an exercise in a context when the parties have entrusted the deciding of the merits of the arbitrators? Even more pointed is the dilemma which arises where the court is asked to grant injunctive relief to prevent a party from doing something which the arbitrators may ultimately decide when they make their award he was in reality always entitled to do. Can the court fairly consider whether to grant an injunction without some assessment on its own behalf of the very question which the arbitrators are appointed to decide; and if the answer is negative - what kind of assessment is this to be? I cannot enter on this question now but it will I believe arise quite soon in many domestic jurisdictions and it could I (think) be a fruitful area for exploration in an international context such as the present.

This leads me to another, and rather different kind of tension: the inevitable discordance between ambition and achievement. Those concerned, as we all are, with enhancing the utility of the arbitral process to the national trading community should perhaps beware of attempting too much. The choice of arbitration, like almost all choices, involves a "trade-off ": something gained and something given up. Parties choose arbitration because they conceive that it will bring a benefit, but the [Page120:] choice carries with it certain disadvantages, notably the inability of the arbitrators to enforce their procedural orders. Those who enter into arbitration agreements must be taken to accept that in some circumstances the arbitral process may be inefficacious. If we try too hard to patch up the process by elaborate contractual provisions or by judicial intervention we may create inflexibility, duplication and a distortion of the true basis of a reference to arbitration. Might it be better simply to live with inefficacy in some respects, including some aspects of interim relief, as the price to pay for a method of dispute resolution which is so far as possible free-standing?

I will next offer some unconnected comments on procedural orders by arbitrators, arising from what has been said earlier today. In the first place, whilst it is true that many, but not all, of such orders may involve a preliminary assessment by the arbitrators of the merits of a dispute [t]hat he will later have to decide by a final award, this much less objectionable than when the court makes a similar order; for the arbitrators have been chosen by all parties to make the final decision whereas the court has not.

I go on to explain why I have just referred to "many but not all" such orders. As it seems to me, procedural orders fall into three broad categories. First, there are orders which reinforce the decision-making process - for example an order for the production of witnesses or documents, or for an expert to make an examination. Second, an order for the protection of the subject matter (or its monetary equivalent) pending the award - such as an order that the subject matter shall be preserved in storage, or that deteriorating goods shall be sold and the proceeds placed on joint account. Third, an order for the preservation of the worth of the arbitrator's ultimate decision on the merits - for example to prevent the defendant from adopting a course of conduct which, if persisted in until award, will make even a favourable award useless to the claimant. It is only in the final category that the court will be called upon to form the kind of preliminary opinion on the merits which might be thought to be a usurpation of the proper functions of the arbitrators.

Various other means of categorisation have been suggested today. I would also propose, as distinct categories - (a) those orders which affect only the parties to the arbitration, and (b) those which affect third parties (such as the freezing of bank accounts, and orders requiring witnesses to give evidence). There is a large difference here. Participation in an arbitration agreement involves consent by the parties to at least some measure aimed at making the chosen dispute-resolution process more effective. But third parties have never consented to this, especially where the arbitration is going on in a foreign country. What they do tacitly consent to is the application of their own local domestic arbitration law by the local court: and if the local law enables the local court to grant interim relief, so be it. But it is altogether more radical to subject a person not party to the arbitration agreement to an order made by an arbitrator to whose appointment he never agreed, and whose participation in the dispute is purely private and voluntary. Thus, if as several speakers have proposed, there ought to be a mechanism for conferring some coercive force on the orders of arbitrators, this should I suggest be confined to orders made against the parties to the reference.

I turn to the question whether such coercive powers should be created. Arbitrators have traditionally relied upon their own moral authority to ensure that their orders should be observed. (I must note a suggestion that arbitrators can compel obedience by making it clear that disobedience may have an adverse effect on the ultimate award. I must respectfully demur. If the merits are such that a party is entitled to win, then it is the arbitrator's duty to decide in his favour however recalcitrant he may have been during the arbitration. But perhaps I have misunderstood the suggestion.)

Nowadays, however, the marked decline in the standard of conduct in contemporary arbitration has increased the risk that the arbitrator's moral authority will not be enough. Should there be some method of attaching the court's sanction to the arbitrator's orders?

How could the orders of arbitrators be given the force of law; how can they be self-enforcing? Is there for example the possibility of international legislation, designed to be incorporated into municipal law, which would enable the orders of arbitral tribunals to be the subject of exequatur? One possibility, which already exists in some systems, is for the arbitrators to make their orders in the form of interim awards. I do not find this attractive. Under some systems it is not an option. Moreover, even if a valid interim award could [Page121:] be made this might often be too slow. It would first be necessary to assemble in country A, the chosen seat of the arbitration, the three members of the tribunal who, almost by definition where institutional arbitration is concerned, are drawn from the furthest corners of the globe and have full diaries. At length the question of interim relief is debated before them by the lawyers of the parties, who come from countries B and C. They make an interim award requiring steps to be taken to preserve property in country D, and the deposing of witnesses residing in country E. By the time the orders are made many weeks have passed. Even if the defendant in country C is willing to obey the order, an assumption which may be excessively naive, steps will have to be taken in countries D and E to put the orders into effect in countries D and E, not only against the defendant but against third parties who have no connection with the arbitration. For this purpose it will be necessary to instruct lawyers in D and E who have no previous knowledge of the matter, and then bring the question before the courts in D and E, who are no doubt already very busy. I have of course chosen a very extreme situation, but even in simpler cases the time scale may well be too long.

Nevertheless, I do not wish to be nihilistic. There are precedents which it would be worthwhile to explore. First, there is the New York Convention. Second, there is the Convention on Jurisdiction and the Enforcement of Judgments, which already extends in some degree to the enforcement of pre-judgment orders made by courts, including some of the types of order which we have discussed today. I suggest that it would be a most valuable initiative if our hosts today, or some kindred body, were to initiate a study, not weighed down by excessively optimistic expectations, to see whether precedents could be developed into an international regime of the mutual recognition of interim measures ordered by arbitrators.

The establishment of a statutory regime of this kind would inevitably be a long process, and as an interim and partial solution consideration might be given to the development of standard rules to be included in arbitration agreements, expressly empowering the arbitrators to order interim measures of protection. Of course, these could not bind the courts, but if an application for similar measures were to be brought before a court with formal power to order them, a judge who might otherwise be hesitant about making an order, for fear of trespassing on the territory of the agreed arbitral method of dispute resolution, would be reassured to know that he was in fact acting with the express consent of the parties.

I have one final comment on the contrast between institutional and ad hoc arbitration. In my opinion no outright comparison of their merits is possible; for some types of dispute one form is preferable, whilst the other is better for a different type. At its best, ad hoc arbitration is flexible, versatile and quick. At its worst, it is ambivalent and weak. At its best, institutional arbitration provides a framework within which parties (especially those who have no great practical experience of arbitration) can know what is expected of them, and on which the parties can fall back as an external and independent resource to solve at least some of the problems which may arise during the arbitration. At its worst it is bureaucratic, cumbersome, unimaginative and slow.

Amongst the advantages of institutional arbitration is its ability to impose some measure of interim relief without recourse to the courts. It seems to me however that its role must be limited by three factors. First, the need for interim relief is often urgent. I shall be very interested to see how the ICC's new body works in practice, and I certainly wish it well, but I will be a little surprised if it can react as quickly to an emergency as a court operating at its best. Second, an arbitral institution has no more right to make orders affecting third parties than the arbitral tribunal itself, so there will be some very important categories of interim relief which will lie outside its powers. Third, there is a point which is easy to overlook namely that the institute is more at risk in making a premature determination of the merits than the arbitral tribunal itself. As already noted, some types of relief require a preliminary and tentative estimate of the strength of the claimant's case. There is nothing in principle objectionable in having the assessment made by the arbitrators themselves, since they are the persons to whom the parties have entrusted the final determination. The position of the institution is different. It is not the arbitrator and is no more entitled to enter into the merits a fraction beyond what is absolutely necessary than is the court itself. So I fear that the interposition of an institution will merely move the problem from one place to another without actually solving it. I hope that I am proved wrong, and believe that in any case the experiment will be well worth trying.