THE AUTONOMY OF ARBITRATION: AUTONOMY A GEOMETRIE VARIABLE

Different arbitral legal theories have different views on the autonomy of arbitration, with which they usually refer to its independence from domestic legal orders. The debate is often heated, but no end is in sight. We submit that one important reason for the difficulty to come to any definite and somewhat universally accepted solution may well be that neither "autonomy" nor "arbifration" are clearly defined terms, and are often used in an opaque and inconsistent manner. In this article, we seek to clarify both notions. As to "autonomy", the focus of the debate is usually on legal autonomy. A solution to the autonomy debate would then require a common understanding of law, which is, however, nonexistent. We therefore introduce, as "neutral" outside tools, the methodology of Niklas Luhmann's systems theory, which observes stable and coherent communication contexts generally, some of which are then qualified as "legal". This frees the autonomy debate from a particular understanding of law. As to "arbitration", we will observe different layers. At a global level, arbifration exists as a phenomenon with a certain intellectual autonomy, but it is not coherent enough to operate as an autonomous legal order. Individual arbifral proceedings, on the other hand, are autonomous in a sociological sense (as interaction systems), but not in legal terms. Rather they depend, and this is our main novel contribution to the autonomy debate, on non-state legal orders at transaction level-lex transactions orders. The latter are autonomous in sociological terms. The extent of their legal autonomy depends on the circumstances of the case. The phenomenon of arbitration overall thus benefits from autonomy a geometric variable at different levels, in multiple forms, and in varying degrees.