Here are my prepared remarks for yesterday’s conference, Law’s Porosities, in Washington. Thank you to Vivian Curran for the kind invitation to participate!

In this talk I am going to give an example of how a breakdown in the idea of territoriality in light of changing technology is causing turbulence at the edge between the common law and civil law worlds. Perhaps it is even an example of porosity. My example comes from the field of civil procedure, and more specifically from the field of service of process. I will start by outlining the traditional territorial model of cross-border service of process. Then I will outline the analogy between service by postal channels and service by electronic mail, explaining at a very high level along the way the basic structure of modern email and some relevant differences between the two methods. Last, I’ll explain the problem service by email poses for the traditional territorial ideas involved in service of process and sketch some of the issues that will have to be overcome if electronic service is to become more universally accepted. I am going to conclude with more questions than answers.

While defending the desirability of electronic service isn’t the point of the talk, I want to say at the beginning that I do think electronic service should have a role. The postal services are reinventing themselves as carriers of the goods we buy from online retailers, but the volume of letters and other first-class mail, at least in the United States, is in a long-term decline. There will always be a place for traditional letters, particularly on important occasions, and because service of process is an important occasion, there will probably always be a place for service of process by mail. But there is no sound reason why electronic transmission of information shouldn’t take its place in this area just as it has in, say, contract law, and as in contract law, it is increasingly anachronistic to require a physical embodiment—a contract with a seal, a physical summons—in order to give a document legal effect.

Today the main international convention governing service of process across the Atlantic is the Hague Service Convention. Concluded in the mid-20th century, the Convention set up a system of central authorities. The idea is that each state designates a central authority to receive requests for service from abroad. When the central authority receives a request, it executes it in accordance with its own local law and then sends a certificate back to the person who made the request for service. But while the central authority system was the Convention’s main innovation, the Convention also preserved the right of a plaintiff to use certain alternate methods of service, as long as the state of destination does not object. Of particular note for our purposes, the Convention “does not interfere with” the right of a plaintiff to serve process by postal channels, provided the state of destination does not object.

Does the Convention allow for service by electronic mail? Not today, in my view. But if it does allow for service by email, it does so by treating email as analogous to postal mail.

Why might a state object to service by postal channels, or by another alternate method? This question can puzzle common lawyers. In the common law world, at least in modern times, we are very lax about who is competent to serve process. Under the Federal Rules of Civil Procedure, for example, any person other than a party is competent to serve process. The reason is that the common law courts have gone very far along the path of distinguishing process, which serves the function of providing notice, from personal jurisdiction, which serves the function of ensuring that the court hearing a case has an adequate relationship to the defendant to make the exercise of jurisdiction fair. There is still some conceptual confusion, and a US court will typically say that it cannot exercise jurisdiction until the defendant has been validly served. But I think the best view, from a common law perspective, treats service almost exclusively as a means of providing appropriate notice. From this perspective it does not much matter who provides the notice, except to the extent there is an issue about proof of service.

But in the civil law world, there is much more attention paid to the issue of who is competent to serve process, because service of process is conceived as an exercise of the sovereign’s power rather than just as provision of notice. Perhaps also the relative prestige that attaches to the profession of huissier, in contrast to the profession of “process server” in the United States, contributes to the civilian view. In any case, it would be improper for me to hire a US process server to fly to France and serve a summons on a French defendant. I am sorry to say that there are examples of US lawyers who should know better doing just such a thing. France has not objected to service by postal channels, though Germany has. So France will allow service on people in France without the involvement of the huissier, though with the involvement of the French postal authorities. But Germany is not so liberal.

The civil law state’s objection to non-standard forms of service is essentially territorial. No one can do certain things on French soil except French authorities. No one can do certain things on German soil except German authorities. A particularly stark example of this approach is found in Article 271 of the Swiss Criminal Code, which provides: “Any person who carries out activities on behalf of a foreign state on Swiss territory without lawful authority, where such activities are the responsibility of a public authority or public official, [or] any person who arries out such activities for a foreign party … is liable to” imprisonment.

I said earlier that if the Convention allows service by email, if at all, only by analogizing email to postal mail. How does the analogy work? Let’s start with a simple past few if any of us experienced. In the early days of the internet, if Amy, in America, had a networked computer and she wanted to send an email to Francois in France on his networked computer, she would simply send the message directly from her computer to his computer using a protocol called SMTP, which incidentally is still the protocol used for email transmission and reception today. By analogy to the postal channel, this early form of email was like sending mail without the postman.

A more modern example: Amy gives her letter to the postman, who then gives it to the French postman, who delivers it to Francois. Or else Amy puts her letter in the mailbox, and the postman then gives the letter to the French postman, who delivers it to Francois’s post office box, and Francois then retrieves it. This begins to approach the architecture of modern email. Today we do not transmit emails directly from point to point: we have email service providers (think Gmail or Yahoo). We send email to our providers, who send it along to the recipient ‘s provider. The recipient then can retrieve the message from the provider using the IMAP protocol. It’s important to note that in both the email example and the postal example, delivery is complete when the message arrives at the mail box or the recipient’s email provider’s server. This is true in a technical sense and a legal sense. If I never take my letters out of the mail box or never check my email, I can’t claim I never received the message. And in the IMAP protocol, the email provider does not transmit the email on to the recipient’s computer. Rather, the recipient reaches out to the server to access the email—to read it or manipulate it in some other way—while the email remains on the server.

There are a few salient differences between the postal channel and email. For one thing, the government is not a party to the commercial email that we all use. For another thing, it is easier given today’s technology to prove receipt of postal mail than it is to prove receipt of an email. But the difference I want to highlight relates to territoriality. In the postal examples, we know where Amy is, we know where Francois is, and we know where the French and American postal authorities and post offices are. The post office is entirely territorial.

Email is not so simple. In the first example we gave, from the early days of email, we do know where Amy and her computer and Francois and his computer are, and the email is transmitted directly from one computer to the other. But modern email makes use of “cloud computing.” By “cloud computing,” I mean the use of datacenters with storage, computing, and other capabilities all around the world. When Amy submits her email to her email provider, she has no way of knowing the physical location of the server to which her computer transmits her message. When her email provider transmits the message to Francois’s email provider, no one has a good way to know the physical location of that server. And when Francois accesses the message on his email provider’s server, again, we have no way to know its physical location.

Now let’s bring these ideas together. I said at the outset that I thought making space for service by email was desirable and that service by email was permissible under the Convention, if at all, only by analogy to service postal channels. So if we are to allow service by email, either we have to amend the Convention or we have to clarify that email is permitted under the provisions of the Convention on postal channels. Either way, we have to ask about how to treat objections by the state where the defendant is present. Does it make sense to allow objections? Does the rationale for civilian objections hold in the case of cloud computing?

If the rationale for allowing objections is territorial, then the answer seems to be no. When Amy sends a document to Francois by post, we know where the document is at all times. It ends up at a post office or a mail box on French territory. This simply isn’t so in the case of email. When the email is finally delivered, it resides on the email provider’s server, but that server can be anywhere. The delivery is complete at that point. If and when Francois reads it, he may be in France, but at a technical level he is reading a document that is stored on the email provider’s server, just as when I read a web page I am reading a document stored elsewhere. If we carry the analogy to its conclusion, it’s as though Francois traveled to the location of the server to read his mail, and then traveled back to France. So if the only idea at play is territoriality, then it seems there is little rationale for allowing objections. The foreigner has done nothing on French or German territory. To be sure, the foreigner has affected the legal rights and obligations of a person on French or German territory, but the Convention does not prohibit methods of service such as service by publication that do the same thing.

How should we go about working towards service by email in the framework of the Convention? Certainly the simpler path is to treat email as part of the postal channel and thus to avoid the need for a renegotiation. But I am not sure this can work, given the power of states to object to service by postal channels and the rationale of that power in the idea of territoriality. The key question is: where is the defendant at the time he is served? If he is in Germany, say, then the German objection to service of process by postal channels governs, whatever his nationality, and indeed, whatever his connection or lack of connection to Germany, aside from mere physical presence at that moment. Suppose I am on vacation in Germany, someone here in Washington sues me, and the plaintiff serves the summons on me by email. I retrieve the email from my email provider’s server and read it while in Germany. Should Germany have a power to object? The plaintiff has done nothing in Germany and may not even have known I was in Germany. I am not a German national and it is difficult to see that Germany has any interest in protecting me.

On the other hand, it is difficult to see a successful negotiation that results in a Convention that continues to permit objections to service by postal channels but does not permit objections to service by email. I say that without extended explanation: it is simply my sense that the more traditional a method of service, the more acceptable it will be to many states.

So as I said at the outset, there are more questions than answers. We will have to begin with considering alternative, non-territorial bases for objections to new electronic methods of service that rely on cloud computing. Perhaps this will lead to a reexamination of the territorial bases for the objections that the Convention permits for existing methods of service, or perhaps to a reexamination of the function of service of process, which as I noted is a good illustration of the difference in perspective between our two legal cultures.