Authenticate the event,
not the autograph.
Anyone can ask whether a signature looks right. We ask whether this book, this page, this ink and this handwriting could all have come together at one real signing. Here is how we find out, in five stages.
“Does this look like the author’s signature?”
“Could this book, page, ink, handwriting and inscription all have existed together, at that moment?”
Resemblance comes last, because resemblance is exactly what a forger works hardest to supply.
Stage I · ChronologyCould it have happened?
Before we look at the handwriting at all, we ask whether this exact copy could ever have been signed by this person.
Read the method in full
- Identify the exact book.
- Publisher, edition and printing, read from the copyright page and its number line, not from the seller’s description. A copyright date is not a printing date: a page reading “© 1996” can be in a copy printed years later.
- Find the earliest date this copy existed.
- A later printing cannot carry an earlier signing. Proofs and advance copies are the exception: they exist before publication.
- Set the signing window.
- From the day this copy was made to the last day the author could sign. A dated inscription narrows it to that date.
- Match the period.
- Signatures change over a lifetime. A signature dated 1997 should look like the author’s signatures from the 1990s, not their later style, which forgers often copy because it’s the one most shown online.
- Test the inscription’s logic.
- Does it mention a book not yet published, a title not yet held, a place that didn’t exist yet? We separate the impossible (strong evidence) from the merely unrecorded (most signings were never documented).
Signed in 2001: Possible. Now the real examination can begin.
Stage II · Physical examinationIs the object what it claims to be?
Then we check that this is real handwriting, written by hand directly onto this page of this book.
Read the method in full
- Rule out reproductions first.
- Under the microscope, printing shows dots or flat ink lying on the surface, rubber stamps leave a halo at the edges, and signing machines (autopens) draw unnaturally even lines with blunt starts and stops. Some publishers print facsimile signatures in special editions, so we identify the edition before calling anything a fake.
- Read how the ink met the paper.
- A real pen sinks into the fibres. Ballpoints leave striations and skips, felt-tips feather and often show through on the back of the leaf, ink pools where the hand slows or turns.
- Look at the back of the page.
- Raking light shows the pressure of the pen: heavier on downstrokes, lighter on connections. Indentation with no matching ink is a sign of tracing.
- Use ultraviolet light.
- UV reveals inks that look the same in daylight but aren’t, erasures and retouching, pencil guidelines, and a page that glows differently from its neighbours because it was added from another copy.
- Confirm the page belongs to the book.
- A genuine signed page can be moved into a more valuable copy. We compare paper, edges, sewing and toning with the leaves around it. A real signature on a transplanted page is not an authentically signed book.
Ink sinks into the paper fibres and feathers along them; the line swells and thins with the pressure of the hand, and every stroke starts and ends in motion.
Stage III · HandwritingDoes the writing behave like the writer’s?
Only now do we compare the handwriting, against the author’s own writing from the same years, looking at how it was written rather than what it looks like.
Read the method in full
- Record the signature before looking at examples.
- We note its shape, proportions, slant, connections and line quality first, so we don’t read into it what we expect to see.
- Use only examples we can trust.
- Comparison is only as good as its examples, and the market is full of forgeries copied from forgeries. We rely on documents of undisputed origin, and treat auction and internet images as context only.
- Compare like with like.
- Examples from the same period, written in books, with a similar pen, and in a similar situation: a hurried signing-line signature for a hurried one, a careful dedication for a careful one.
- Learn the writer’s natural range.
- Size and slant wander; stroke order, proportions and certain letter constructions barely change. Those stable habits are what a forger fails to notice.
- Treat the inscription as part of the event.
- A dedication and its signature should be one act, in one hand and one ink. Forgers practise the famous signature, not the ordinary words around it, and sometimes add an inscription to a genuine signature to raise its value.
Six genuine signatures by the same writer. Look closely: none of them is identical to another.
Stage IV · ControlsCan anything else explain it better?
Next we try to prove ourselves wrong, testing every other explanation that could account for what we see.
Read the method in full
- Tracing.
- Slow, even pressure, hesitation on curves, blunt ends, guidelines under the ink, and too perfect a fit to one known example.
- Practised copies.
- A skilled forger can write fluently, but usually captures the famous outline and misses the quiet habits, or uses the style of the wrong period.
- Honest substitutes.
- Not every signature that isn’t the author’s is a forgery: secretaries and family sign on authors’ behalf, publishers print facsimiles, and signed bookplates are genuine but weren’t signed in the book. Each is described for what it is.
- Bias.
- We examine the item before the examples, ignore the price and the story, and have a second examiner review it independently.
- Weight, not count.
- A slant or ink colour that matches means almost nothing; a rare habit that matches means a lot; one fundamental difference can outweigh many similarities. For each observation we ask: is this more expected if the signature is genuine, or if it isn’t?
Add observations to see how much each one weighs.
Stage V · OpinionWhat does all the evidence support?
Finally we weigh everything together and state an opinion only as strong as the evidence allows.
Read the method in full
- Five streams of evidence.
- Chronology, the book, the ink, the handwriting and the internal logic of the inscription. They should all point the same way; if any conflicts, the conflict is explained before anything is certified.
- Impossibility decides.
- A printing made after the author died, a mechanically reproduced signature, a page from another copy or a near-exact copy of a known example each settle the question on their own.
- A graded opinion.
- Our opinion is stated on a graded scale, not as a bare yes or no, and “inconclusive” is an honest result we’re willing to give.
- Two examiners.
- At our highest level, Black Label, two experts examine the item independently and certify it only when their conclusion is unanimous.
- A permanent record.
- Every certificate is entered in this register with the photographs taken during the examination, so anyone can check it, at any time.
- 1Genuine
- 2Highly probably genuine
- 3Probably genuine
- 4Indications of genuineness
- 5No conclusion
- 6Indications against
- 7Probably not genuine
- 8Highly probably not genuine
- 9Not genuine
On every Veritado certificate
What we do
- Identify the exact edition and printing from the book itself
- Examine the signature under the microscope
- Examine the page and ink under ultraviolet light
- Check the back of the signed page
- Compare with verified examples from the same period
- Photograph the item for the public record
- Black Label: two independent examiners, unanimous
What we don’t rely on
- Receipts and invoices
- Other certificates of authenticity
- Photographs of the signing
- Dealer or auction descriptions
- Stories about where an item came from
These can be true and still prove nothing about the signature in front of us. We judge only what the object itself and reliable comparison material can show.
One coherent event
the right book+the right page+the right period+the right materials+the right handwriting+the right historical logic
When they all agree, the signature is certified. When any one of them doesn’t, it isn’t.
Check a certificate