Email history (presentation at Connections 2011)

The following is a short essay I wrote on email history for Connections 2011 in Milwaukee. Since email history is still a nascent field, there’s much more to say, but here’s a rough cut.

When I first wrote this paper I wanted to accomplish two things: 

1) I wanted to write the history of email. Preliminary research indicated that no historiography of email existed. On the Web there are resources, many hagiographical. There is Ian R. Hardy’s MA thesis from 1996, but it has historiographical deficiencies. A second, and perhaps more authoritative source for the history of email is Craig Partridge’s article “The Technical Development of Internet Email”. While Partridge offers detail and historiographic sophistication not available in Hardy’s MA thesis, it suffers from a  fatal flaw. Partridge begins his history at 1971, with RFC 196 and Ray Tomlinson’s SNDMSG update for the TENEX operating system. While this was certainly a watershed moment for email, since it created a sophisticated, dedicated email technology capable of transferring mail over the ARPANET, it was hardly the first email system. To tell the full story of email you must go back to the mid-to-late 1960s, especially Douglas Engelbart’s oN-Line System. Today I am going to touch upon some of the connections between these developments in the 1960s to the networked email systems that developed in the early 1970s.

2) This brings me to the second goal of this paper. I wanted to offer a materialist presentation of email, in the way DeLanda has accomplished in “War in the Age of Intelligent Machines” or “1000 Years of Non-Linear History”. To do so, I must expose the contingencies of email as a socio-technical apparatus. Without telling the development of email in the 1960s many contingent practices and technologies are obfuscated. In particular, there are two such contingencies that are politically latent and important to the development of email technology.

The first is the role of the military, as the primary sponsor of the ARPANET, which eventually came to be the Internet. Disentangling the development of the ARPANET and email is complicated, since the technologies co-developed, and share many origins. Email played an accidental, but central role in satisfying key motivations for the US military’s investment in ARPANET, which at the time was undergoing a shift to a new command and control structure.

The second contingency is the 1950s and ‘60s epistilatory traditions. Although it is rarely discussed, the development of email was considered antagonistic to the “proper” goal of ARPANET, which was designed for resource sharing through the military. Not only does the discourse of email pick up on epistilatory metaphors (mail, letters, envelopes, and to:, from:, reply:, etc.), I am arguing that it shares similar salient political goals. These metaphors are examples of what N. Katherine Hayles calls “skeuomorphs” and “seriation charts” and they show the tight connection between letter writing and the development of email. When I first saw Thierry Bardini’s diagram of the development of the “personal interface” (from his book “Bootstrapping”) I found it illuminating for a number of reasons. It covers roughly two decades of complex and co-articulated technologies, but right through the middle of the chart is a line that cuts through all technologies. Although Bardini only mentions it in passing, I think the symbolic weight of the date cannot be underplayed. Cutting across all technologies is 1968, a clear reference to the “May ’68 events” in France. While I don’t have the time to articulate the way the ’68 events impacted American culture and outlook (and indeed, it’s an area of future research, since I think it is undeniable that ’68 was a moment of rupture for the development of computing technologies). So, before the ’68 events there was a groundswell of political and personal engagement in what might be described as “personal augmentation”. This description is rather too banal, however, since the phenomena was more like Tom Wolfe’s “Electric Kool-Aid Acid Test”, with LSD and other psychedelic use ranging from disciplined and measured to the “Day–glo”. In the broadest way, the relationship between psychedelics and the development of computing technologies has already been discussed in John Markoff’s book “What the Dormouse Said”. The specific relationship I want to discuss is from the decade prior. Although I won’t elaborate on it here, Douglas Engelbart and his team at the Stanford Research Institute used psychedelics to further their technological goals. But, the decade prior, and the intellectual and cultural driving force for the “acid test” generation, was the beat poets of the late 1950s. Not only did these poets use psychedelics and other drugs to further personal and political ends, they conceptualized an epistilatory tradition that was to see open letter writing as politically meaningful by invoking an “open secret”.

The initial DARPA requirement for the ARPANET was to provide networking capabilities for resource sharing.   A pioneering spirit for the ARPANET was Joseph Licklider who, in 1962  argued that computers could be used for more than resource sharing. Licklider wanted to “improve man–machine interaction in teaching and learning, in planning and design, and in visualizing the internal processes of computers”. Licklider later argued, “I wanted interactive computing, I wanted time-sharing. I wanted themes like: computers are as much for communication as they are for calculation.”  Time–sharing for communication, not resource sharing, soon became the new computing prerogative.

While Licklider was laying the groundwork at DARPA for what would eventually become the ARPANET (which email would function across), Douglas Engelbart was developing the oN Line System at the Augmentation Research Center in the Stanford Research Institute. NLS was a very early implementation of networked computers that, in 1971, also joined the ARPANET. Before the existence of the ARPANET, NLS was a system of networked communication that, unlike later ARPANET implementations used closed–circuit television to display terminals remotely.  The Journal and Mail subsystems made NLS a unique and important precursor to email. The Journal subsystem was conceived in 1966 for the purposes of keeping a “log” of events, and performing a document–oriented communication system, described as “direct distribution”. Direct distribution could send documents (memos, messages, data records, etc.) directly to invited participants through the use of a personal IDENT code. Documents were sent to a “mail box” and marked with a status, such as “For Action” or “For Information”. The Journal began a full five years before the accepted “official” start–date of email,  and yet the Journal remained, co–developing in similar and different ways alongside other systems of email. Many parts of the system were shared with proto–email systems, such as the IDENT codes, directory lookup and mailing lists (mirroring similar functionality developed later in email).

A parallel story begins in 1961, when the development of PLATO II and CTSS introduced time–shared computing. Time–sharing quickly became popular and through the 1960s it was common to pass notes to other users by leaving a file for another user by placing it in a common directory. Tom Van Vleck suggests that it was common to title the file left in the common directory with a person’s name, such as “to tom”.  The first system to formalize a mail command occurred on CTSS at MIT. Between December 1964 and January 1965 the undated Programming Staff Note 39 for CTSS was written and distributed.  In February 1965 Van Vleck joined the programming staff at MIT, along with Noel Morris shortly thereafter.

Over the spring of 1965 Van Vleck and Morris read Programming Staff Note 39 and on the weekend of July 4th, 1965 they implemented the MAIL subsystem for CTSS. At some point in 1969 Vleck re–implemented MAIL for the Multics time–sharing operating system. These systems, however, were not networked.

In yet another parallel, but slightly later development, on April 16, 1971 RFC 114 was published to specify the File Transfer Protocol (FTP). Almost immediately, FTP was used to send email across a network using a mechanism very similar to the time–sharing technique. FTP relied on the early HOST protocols developed for the ARPANET–connected computers and if a user wanted to send email to a user of another networked computer he or she would log into the remote computer and leave a file for the user, just as the time–sharing users did previously. 

The most significant advance in networked email came with the development in late 1971 by Tomlinson when he added network capabilities to SNDMSG and READMAIL for the TENEX system on DEC’s PDP–10 machine. Previously, SNDMSG was used to send local messages, or even used to send local messages from a remote Telnet connection.

N. Katherine Hayles has imported from archaeology two useful concepts for understanding technological development: skeuomorphs and seriation charts. As Hayles describes it, a skeuomorph is “is a design feature, no longer functional in itself, that refers back to an avatar that was functional at an earlier time.”  Further, skeuomorphs visibly testify to the social or psychological necessity for innovation to be tempered by replication. Like anachronisms, their pejorative first cousins, skeuomorphs are not unusual. Seriation charts are similar in that they “depict… changes in an artifact’s attributes [that] reveal patterns of overlapping innovation and replication.”  A skeuomorph is a physical artefact testifying to an earlier design requirement (Hayles gives the example of the fake stitching on her car’s vinyl dashboard), and a seriation chart is the dynamic morphology of the artefact, or, “overlapping innovation and replication” when a seemingly necessary design arises out of contingency. 

For email, many early design decisions seem obligatory, but highly contingent: for example, the SOML command maintained appearances of direct “instant messaging”. A pervasive seriation is the reoccurring use of the keyword syntax, from the NLS Journal. The “keyword” syntax points to the cognitive and political challenges regarding the use of email—its widespread use as a personal communication system but its funding as a corporate or research memo and document system. The contingency is that research requires categorization for effective information retrieval (or so the story goes), yet categorization hardly makes sense for personal communication. Other skeuomorphs signal a pre–computer era, such as cc: (carbon–copy), bcc: (blind carbon–copy), POSTMASTER as a reserved name, and the HELO command for initiating a new MAIL connection.

While it is commonly assumed that computing technologies are immaterial, or immaterializing, in the case of email a more useful metaphor might be the postcard. Indeed, for email, header syntax is always exposed, even in the case of encrypted messages, so unlike an anonymous letter (or even an enveloped letter), email is more mass than personal communication. As Godard compares film to “the sending of 25 postcards per second,”  email is a kind of rapid postcard (fast, but material, and open). As an electronic postcard, email might offer an updated version of the Beat Poets’ project of “open secrecy”. Said to “declassify the secrets of the human body and soul” open secrecy was associated with the the Romantic idealization of spontaneity, the letter promised to extend the originally oral, intimate, and mutual confessions of the early Beat circle. Controversially, perhaps truly secret communication is more like terrorism (by the state or otherwise), whereas an open secret has political relevancy and potency.

Derrida recognized that postcards are especially open–ended communication media, since, without a stamp the postcard will never reach its destination, and words that never arrive are “rendered unreliable.”  Yet, truth can transcend media, since it does not require a stamp to arrive. Email, like the postcard, is a media with “double potential”: it can transmit a message or it can interrupt a message (if the card never reaches its destination). This is unlike the telephone or instant messaging, which is interactive and supplies contextual clues to the sender when the message has been or has not been received. Camelia Elias writes, “when the message is original, it has a textual structure; when the message is potential, it has hermeneutic structure.”

In closing, I will just mention there are many areas of future research: first: solidifying the connections between the 1960s political and cultural milieu and the technical development, second: articulating the role of DARPA and the military in email’s history. And, there are probably many more, but I’ll stop here. Thank you.