A real pen, a mechanical arm, and a plastic template

The autopen is the strangest problem in autograph collecting, because what it makes is not a forgery. It is the person's own signature, produced by their own machine, at their own instruction.

It just isn't their hand. And it has been in general use since the late 1940s, which means a lot of what is sitting in frames right now was never touched by the person whose name is on it.

How the machine works

It is more mechanical than most people expect.

Someone signs their name a few times. A company cuts those signatures into a template, or matrix. In the Apollo era that was a large Plexiglas spiral with indentations in its surface, loaded onto a turntable beneath the desktop. A mechanical arm holding an ordinary pen, pencil or marker follows the groove, and writes on whatever you place under it.

The model NASA used most in the Apollo years was the Autopen Model 80, made by the Damilic Corporation of Rockville, Maryland. It looked like a small drafting table with a writing arm on one side and space for a stack of photographs on the other.

Matrices swap in and out easily, so one machine can sign for several people, and one person can have several matrices with slightly different versions of their signature. That last part matters later.

It started in a torpedo factory

The idea is much older than the machine. In 1803 an inventor called John Isaac Hawkins patented a device he named the polygraph, nothing to do with lie detectors, which reproduced handwriting mechanically. Thomas Jefferson took an interest and is said to have used one frequently. Hawkins also designed the upright piano.

The modern machine comes from the Second World War. Robert De Shazo Jr was twenty-one and working at the Naval Torpedo Factory in Alexandria, Virginia in 1941 when he encountered a device others had built and judged not commercially viable. After the war he bought the company, reworked the design, and named it the International Autopen Company.

His first order came from the Secretary of the Navy. It spread through the federal government from there.

By 1958 a machine could produce three thousand signatures in an eight-hour day. In 1983 De Shazo estimated around five hundred autopens were in use in Washington, and made the observation that within four or five square blocks there were more people who needed them than anywhere else in the world.

The part that will bother you

NASA had a genuine problem. The Apollo astronauts could not sign every autograph request and also do their jobs, and there were a great many requests. So the agency chose between refusing everybody and sending machine-signed photographs.

It sent machine-signed photographs. That is defensible, and most recipients were perfectly happy.

Here is the part that is harder to sit with. NASA gave autopenned photographs to the astronauts themselves, to hand out in person.

So somebody meets an astronaut at an event, is handed a signed picture, and goes home believing they watched him sign it or that he signed it for them personally. They tell their family that. The family inherits the photograph and the story together.

It goes further. Some astronauts ended up with autopenned signatures of their own crewmates in their personal collections, without knowing it.

The reassurance that is worth nothing

If you own a crew-signed photograph, you have probably been told, or have told yourself, that it must be genuine because each signature is in a different ink.

It is the most natural inference in the world. One machine, one pen, one colour. Three colours, three men.

It proves nothing at all.

Each astronaut generally had his own machine, and each machine could be loaded with whatever pen was to hand. A crew-signed picture was assembled by carrying it from office to office, collecting a machine signature at each stop, in whatever ink that machine happened to hold that week.

Different inks on a multi-signed item are exactly what you would expect from autopens. They are not evidence of anything.

Who else has been doing this

Presidents, comprehensively. Kennedy began using secretaries around 1955 and autopens heavily during his campaign from about 1959, though in the White House he signed most things himself. Ford used mainly autopens. Carter used both and leaned toward secretarials, his executive secretary being notably good at his handwriting. Reagan used the machine extensively as president and his patterns turn up constantly.

Every president since has used one to some degree. Truman used one. Lyndon Johnson let his be photographed for a National Enquirer article in 1968, which is a level of candour the subject rarely gets. And in 2005 the Justice Department advised George W. Bush that he could sign a bill by autopen. He declined, and went and signed it himself.

Beyond that: every member of Congress has access to a machine. Supreme Court justices. Corporate executives. Actors, sometimes through fan-mail companies that keep a matrix on file for each client.

The machines can also write short inscriptions. "Best wishes" is within their range, and so are small drawings. The advice from the autopen companies was to use a separate matrix for the inscription so the two sit at slightly different angles, which looks more natural.

Read that sentence again. The industry building these machines was advising customers on how to make the output look less machine-made.

How to catch one

The whole detection method rests on a single fact.

No human being signs their name the same way twice. Speed, pressure, pen angle, mood, what you are leaning on. Two genuine signatures by the same person are always slightly different.

A machine following a groove produces an identical result every time.

So the definitive test is superimposition. Hold two examples up to a window, or overlay them on a light table or on clear acetate. If they match perfectly, in proportion and letter shape, one of them came from a machine and so did the other.

Three supporting tells, none conclusive alone:

The pressure is uniform. Human writing has light upstrokes and heavy downstrokes because your hand accelerates and slows. A machine applies constant pressure, which gives the line a flat, even look.

The ink sits heavily. The machine writes more slowly than a person, so ink has longer to soak into the paper.

And it starts and stops oddly. Look for small blobs where the pen came to rest, and for a line that begins with no lead-in stroke.

One genuine caution: a non-match does not prove a signature is real, because the same person may have had several matrices. Failing to find a twin means you have not found the right pattern yet.

Where certification comes in

This is a category where third-party authentication genuinely earns its fee.

PSA, JSA and Beckett all identify autopen signatures, and they hold reference material a private collector does not: catalogued patterns, accumulated examples, and enough throughput to have seen the same matrix many times. Matching a suspect signature against a known pattern is exactly the kind of work that benefits from volume.

The problem is what never reaches them.

Most autopenned material in circulation arrived free, through the post, decades ago, from an agency or an office that answered fan mail. Nobody paid for it, so nobody submitted it. It sits in albums and frames and estate sales carrying a story about how it was obtained, and the story is usually sincere and usually wrong.

Which is the practical advice. If a signature came through the mail from a public figure, an agency or a fan club, treat it as a machine signature until something establishes otherwise. That is not cynicism. It is what the odds actually are.

The thing worth sitting with

An autograph is supposed to be evidence that two people were briefly in the same place, and that one of them made a mark.

The autopen removes the person and keeps everything else: the real pen, the real ink, the real signature, the real paper. What is missing is the only part that was ever the point.

And unlike a forgery, nobody involved was doing anything wrong. The signer authorised it, the office sent it, the recipient was delighted. It is a deception with no deceiver, which is what makes it so hard to argue about and so common in collections.

A forensic science paper studying these machines in the 1970s reached for a phrase that has stuck with me: genuine forgeries. That is exactly what they are, and there is no better way to say it.

This index tracks certified autographs by signer, item type, certifier and grade, and shows what is listed and what has verifiably sold. You can read how we work the numbers out, and there are companion pieces on what a certificate actually claims, the FBI investigation that created the modern authentication industry, and the bat boys who signed for the stars.

Sources