Ownership of Embryos, and Mead, in Roman Law
It’s good to see Roman law help solve a modern problem, and when the problem is especially modern, it’s especially good. Two scholars from England have recently published a piece on rights to embryos, regulated substantially by the Human Fertilisation and Embryology Act 1990, but not substantially enough to bar the help of Roman law. Their basic argument is that treating the embryos as property creates a useful foundation for legal decision making under the HFEA. Here the Roman law helps because, in their view, it supports co-ownership in the embryo, and this turns out to be helpful in certain disputes over their use. (M. Cornell and T. Baron, ‘The law and ethics of a property rights approach to frozen embryo disputes’, 2023 Legal Studies 1-20.)
Cornell and Baron focus on the famous case Evans vs. Amicus Healthcare Ltd but particularly the opinion of the High Court, [2003] EWHC 2161 (Fam), [2004] 2 WLR 713. A woman agreed to the freezing of her embryos against her better judgment – and the assurances of her partner (and contributor) that he would stand by her. He didn’t, and when she sought to use the embryos, he asserted his statutory right to block her from doing so. Her promissory estoppel claim failed. Cornell and Baron think the woman would have benefitted more from proprietary estoppel, which of course requires property in the embryo. Treating the partner as an owner of the embryo in this particular context might prevent him from evading his (otherwise non-contractual) promise.
Roman law supplies the property right, the authors say. The basic idea is that you have two contributors to the embryos, and if you treat them as potential property owners, then you have a situation which maps reasonably well onto the famous Roman ‘mingling problems’: the property of two persons are, by some event, brought together into a unity, creating an ownership dispute. In Roman law, legal and practical considerations determined whether a mingling would be classed either as fusion (confusio), specification (specificatio), or accession (accessio), and of course the choice determined the winner.
The authors favour fusion: this is where the property of two persons are utterly fused – for example the mingling of two liquids – but still identifiable in their former state. In Roman law a fusion brings co-ownership of the final product, a salubrious result that turns neither party away. In a case like Evans, both partners would assume certain rights and responsibilities of ownership in a dispute over the use of embryos.
Naturally much of this depends on what exactly is regarded as the object of ownership, and here we might ask whether Roman property law, accustomed to dealing with grapes and river banks, could deal with the complexity of a living embryo. The answer is yes it could, and without difficulty. The problem is actually the reverse: how to determine the proper underlying object of ownership within a very permissive ownership framework. This is the problem that lingers in the modern cases that have adopted some form of Roman jurisprudence, cases that Cornell and Baron ultimately rely on.
A Scottish case from some years ago addressed the Roman ‘mingling’ of living things: Kinloch Damph Ltd v Nordvik Salmon Farms Ltd & Ors [1999] ScotCS 162 (30 June 1999). There a salmon farmer claimed to own, by specification, salmon that he had raised from smolts belonging to another. The salmon would be a ‘new thing’. It wasn’t an argument likely to prevail, not for the reasons given by the judge (in effect, that Roman law didn’t contemplate the specification of living things), but because the smolts owner was perfectly able to describe his property, and any Roman judge would know immediately that those smolts continued to exist in the mature salmon. (See my ‘Postscript on nova species and Kinloch Damph Ltd. v. Nordvik Salmon Farms Ltd.’)
The lesson here is that the Romans were utterly practical when they sued for property. If a claimant could describe the property he wanted back, then he got a shot at retrieving it from his opponent. The typical mingling problem (in contrast to the salmon case) is simply one in which the property has been transformed beyond description. If its identity had been extinguished (specification, accession), then there was nothing for the claimant to claim back. But if the object were still somehow within the claimant’s powers of description (fusion, commixtion), then he prevailed.
This is what makes fusion such an interesting choice for Cornell and Baron. Fusion sits at the margins of the Roman mingling problems: unlike the usual problems, no property is truly extinguished in the act of mingling. But its boundaries are fuzzy. At one end, there’s a simple mingling where the parties can still describe their former (unmingled) property. This is fusion, and it produces co-ownership. At the other end, the mingling has effected a profound transformation in the former property and the power of description is lost. This is not fusion.
Between these two boundaries, where there’s a mingling and the transformation is less than profound, things get complicated, and this is where the specific object of ownership – or better, what we choose to perceive as the object – matters very much.
The Romans discussed this problem in the context of mead (mulsum): honey belonging to A is mingled with wine belonging to B. If an embryo is the product of Roman fusion, then discussion about ownership starts with these mead cases.
The basic dispute was how far the mead had proceeded along the path of transformation. Did ‘each material remains, despite the mixing’ (utraque materia etsi confusa manet tamen), or did it ‘fail to keep its original character’ (suam speciem pristinam non continet)? (see Justinian’s Digest 6.1.5.1). The abstractions here are basically after-the-fact rationalisations for what was already happening in court. I.e., judges always expected to see the object as described, and the jurists, coming late to the discussion, supplied the Latin. But the point is made: fusion requires special attention to the materials, their presence in the final product, and the overall character of the final product.
This has to be part of the discussion on the ownership of embryos, if fusion is the right path. One stubborn problem is that embryos present much more choice of materials than mead. Is a person claiming ownership of physical matter, or some deeper properties of that matter? In other words, the sperm and egg (the gametes) or some component of the gametes that remain in the embryo? And which component? Cornell and Baron wonder if the transformation from gamete to embryo has obscured the gamete too much. They suggest instead that ‘genetic information’ should be the foundation for a claim, because it obviously remains in the embryo, and because it supports the principle of reproductive autonomy.
The Roman solution is simpler than this, but slightly unnerving. If ownership claims are based on viable descriptions of property, then a contributor claims what he can describe. We should probably admit that the sperm and egg ‘fail to keep their original character’ after fertilisation and are undescribable. So we dive deeper, and wherever we can identify the features of an embryo and attribute them to the respective partner, there is a potential property right in the embryo. And over time, the more acutely we learn to identify those features, the more ‘each material remains’ and the stronger the property right becomes. Tomorrow’s answer may be different from today’s.
In the end it’s really our perception of property that matters, rather than the complexity of the subject or the limits of the law.
