Pre-1858 English Wills: What They Prove, and Who They Leave Out

The will is three pages of dense secretary hand, and by the second reading you have the shape of it. A yeoman leaves his house to his eldest son, twenty pounds apiece to two daughters, and a bequest to a younger son described as being overseas. Four children, named, in the testator's own order of importance. It is the commonest arrangement you will meet in an eighteenth-century English will.

The parish register has seven baptisms to the same couple, and no burials for three of them.

The temptation is to decide that the register is wrong, or that the will is incomplete, or that three children died somewhere unrecorded. Usually none of those is what happened. A will was never a census of a man's family. It was a legal instrument settling what he still owed, to whom, at the moment he sat down to write, and English law gave him no reason at all to mention anybody he had already dealt with.

The short answer. A pre-1858 English will proves that the people it names were alive, or recently dead, in a specific relationship to the testator on the date it was written, and that a church court accepted it as his. It does not prove how many children he had, that the ones it omits were dead, or that the handwriting you are looking at is his. Most of the missing children are missing because of advancement, the practice of settling a child's share early, which the Statute of Distributions of 1670 names explicitly as a thing the courts had to account for. Read the document as a statement of outstanding obligations and it becomes far more useful than reading it as a family list.

The document you are reading is almost certainly a clerk's copy

Start with the physical thing, because the mistake here is silent and common.

The National Archives holds the records of the Prerogative Court of Canterbury, and the series nearly everyone uses is PROB 11: over one million registered copies of wills proved in that court between 1384 and 1858. The word registered is doing the work, and the archive spells out what it means. These wills, it says, "are all registered copy wills. They are the copies of the original probates written into volumes by clerks at the church courts." It adds, flatly, that you will not find any original wills in the collection.

The originals, where they survive, are in a different series, PROB 10, which is not online. The archive describes those as what the executor brought to court to be proved, and is careful to say that they are either the will as drawn up and signed by the testator or a copy made before the court's notary. So even PROB 10 is not a guaranteed autograph. What the archive does promise about it is worth the trip: though the originals hold the same details as the registered copies, it says, they can be easier to read.

Two consequences follow immediately and neither is obvious from the image on your screen.

The hand is the clerk's, not your ancestor's. It is tempting to save a PROB 11 image as a specimen of an ancestor's handwriting, or to read the name at the foot as his signature. It is a specimen of a court clerk's handwriting, written into a register after the testator was dead, and every name in it has been through one more copying stage than you thought. If you are reconciling a spelling against a parish register, that extra stage is exactly where the spelling moved.

Errors of transcription are baked in. The clerk was copying at speed from a document that was frequently in a worse hand than his own. A reading you cannot check against the original is a reading with an unquantified error rate, which is the same problem that surfaces whenever a printed transcript stands in for an image, as it does in the Scottish kirk session minutes where a published 1902 volume spells the same woman's name two different ways inside one case.

The National Archives does publish a small set of specimen wills free of charge, including annotated ones for Jane Austen and for a testator named Susanna Smith, specifically so you can see the standard parts of a will and the styles of writing used across the centuries. Reading those before you buy a stranger's will is fifteen minutes very well spent.

A will lists obligations, not offspring

Here is the folk rule you will meet in every forum thread on this subject: married children were left out of wills because they had already received their portion. It is repeated with total confidence and it is not a rule.

Elizabeth Shown Mills's Evidence Explained site was asked exactly this question in June 2012, by a reader who had a will from 1850 naming only the single children. The answer, posted under the site's own name, opens: "Ann, if this was a custom or a law, it is one new to EE." The site goes on to give the real mechanism, and it is not about marriage at all. Sometimes, it says, testators who had already provided for one or more offspring did not include them in the will, so bequests went only to those who had not received an advance. The common instance it names is a testator with multiple marriages, who might settle with the children of a first wife before remarrying and then, in its phrase, "start life fresh" with the new wife.

That practice has a legal name, and the name is why you can trust that it was widespread rather than anecdotal. The Statute of Distributions of 1670, which governs what happens when somebody dies without a will, contains an express proviso about advancement by portion. Where a child had already been advanced by the parent, the statute required that advance to be brought into account against the share the court would otherwise allot, so that the estate came out equal across all the children as near as could be estimated. Parliament does not legislate a correction for a thing that nobody does. The statute is documentary proof that settling a child early was normal English practice by the late seventeenth century, and that everybody involved understood it to be an advance on inheritance rather than a gift.

Evidence Explained adds the detail that saves you, where it happened: careful testators, or those with a competent attorney, would still name a child who had been provided for and leave them a token, a shilling or some similar nominal sum, precisely so the omission could not be read as an oversight. So a shilling in a will is not an insult. It is usually a receipt.

What this changes in practice. An absent child is a research lead, not a death. The question to ask is not "when did this child die" but "what did this child already get, and where is the document that gave it to them." That document is typically a marriage settlement, a deed of gift, a bond, or a conveyance, and it sits in a completely different record group from the will.

The witnesses are the one group who are probably not family

This is the part of a will most researchers skim, and the law makes it the most interpretable part of the page.

Before 1838, the governing rule for anything touching land was the Statute of Frauds of 1677. Its fifth section required that all devises of land be in writing, signed by the party devising or by another person in his presence and by his express direction, and be "attested and subscribed in the presence of the said Devisor by three or fower credible Witnesses or else they shall be utterly void and of none effect." Three witnesses, watching the testator, signing in front of him.

The Wills Act 1837 replaced that with the rule that is still broadly recognisable today. From 1 January 1838, section nine required a will to be in writing and signed at the foot or end by the testator, or by someone else in his presence and by his direction, with that signature made or acknowledged in the presence of two or more witnesses present at the same time, who then attest and subscribe in the presence of the testator. The Act adds, usefully for anyone puzzling over a sparse attestation clause, that no particular form of attestation is necessary.

Now the detail that makes witnesses worth reading. Section fifteen of the same Act provides that if a person attests a will under which they, or their husband or wife, receive any beneficial gift, that gift is "utterly null and void" as far as they are concerned, though they remain competent to prove the will. In other words, from 1838 a beneficiary who signed as a witness forfeited their legacy.

So the witness list is systematically not the people who inherited. It is neighbours, tenants, the attorney's clerk, the parish officer, the literate man from the next farm. That is a FAN club list handed to you by statute, and it is the raw material of FAN club research: a small set of named, located associates who were physically in the room with your ancestor on a datable day, none of whom stood to gain by being there. If you are trying to place a family in a community rather than on a pedigree, three witness names are often worth more than the bequests.

One more thing section fifteen tells you. If you find a legatee who also signed as a witness on a post-1838 will, you have found either an error by whoever drew the will, or a document drawn before the Act by someone working from old habits. Either is a signal about how the will was produced.

Which court proved it tells you where the property was

England and Wales had no single probate authority before 1858. The National Archives counts more than two hundred church courts dealing with probate, arranged in a hierarchy: the Prerogative Court of Canterbury at the top, the Prerogative Court of York beside it, then bishops' diocesan courts, then archdeacons' courts, then the peculiars administered by deans and chapters.

Which one you were sent to was not a matter of status or preference. It was determined by where the property sat. The National Archives sets out the working rule: personal property in a single archdeaconry or peculiar meant the archdeaconry or peculiar court; property spread across more than one archdeaconry inside a diocese meant the bishop's consistory or commissary court; and property worth more than five pounds in more than one diocese inside a province meant that province's Prerogative Court. Property worth more than five pounds in both provinces required two separate grants.

Read that backwards and the court is evidence. A man whose will went to the PCC held property in at least two dioceses, or died abroad, or died at sea. A man whose will went to an archdeacon's court almost certainly held everything in one small area. The Prerogative Court of Canterbury is often described as the court of the wealthy, and it started out that way, but the National Archives is careful to note that over time a great many less prosperous people used it, and that sailors and soldiers, who feared dying abroad, are very well represented. By the late 1850s the PCC was proving roughly forty per cent of all wills.

This is also the answer to the most common failure in English probate research, which is searching the PCC, finding nothing, and concluding there was no will. There were two hundred other courts, and their records are in county record offices, not at Kew.

A woman's will is a statement about her marital status before it is anything else

Under English law a married woman could not make a will freely. The National Archives puts it plainly in its own guidance: wives needed their husband's permission to make a will, with the result that most female testators were spinsters or widows.

The Wills Act 1837 did not change this. Section eight preserved the restriction explicitly, providing that no will made by a married woman should be valid except such a will as she might have made before the Act passed. The general reform of a married woman's property rights came decades later.

Section eighteen of the same Act adds the rule that catches people out in the other direction: every will made by a man or a woman is revoked by his or her marriage, with a narrow exception for wills exercising a power of appointment. A will followed by a marriage and no new will is a will that died with the wedding, and the estate went to intestacy.

The practical reading. A woman's will, standing alone, is decent evidence that she was single or widowed when she wrote it. A will by a married woman is unusual enough that it is worth asking what was going on: a marriage settlement, a separate estate held in trust, or a testator who was in fact widowed and whose "wife of" description is a clerk's shorthand.

The dates bracket a death, they do not give one

A probate record carries two dates, and neither of them is a date of death.

The date on the will is the date of signing. The date of probate is the date the court accepted it. The death sits somewhere between the two, and the gap can be weeks or it can be decades. The National Archives warns searchers of its PCC index about precisely this, because the index year is the year of probate, and the will itself may date from many years earlier.

The Wills Act 1837 adds a subtlety worth knowing for anything written after 1837. Section twenty-four provides that every will is construed, with reference to the estate comprised in it, to speak and take effect as if it had been executed immediately before the testator's death, unless the will says otherwise. That is a rule about property rather than people, but it has a genealogical edge: a bequest of "all my lands in the parish" covers land bought after the will was written, so the property described in a will is not a snapshot of what the testator owned on the day he signed it.

Section thirty-three is the one that reaches the family. Where a gift is made to a child or other issue of the testator, and that child dies before the testator but leaves issue who are alive when the testator dies, the gift does not lapse. It takes effect as if the child had died immediately after the testator. So from 1838, a will can name a child who was already in the ground, because the law kept their legacy alive for their own children. Before the Act, that gift would simply have failed, and a child who predeceased their parent usually disappears from the will altogether. The same calendar care that applies to reading dates in old records applies here: work out which side of 1 January 1838 your will falls on before you infer anything from a name that is present or absent, because section thirty-four of the Act states that it does not extend to any will made before that date.

Who could not inherit tells you why someone is missing

Some absences are not choices at all. They are common law.

FamilySearch's dictionary of legal terms sets out the position before the inheritance rules were materially changed in 1834, and three of its statements explain gaps that researchers routinely read as family estrangement. A person attainted of treason or felony forfeited his inheritance, a rule that stood until 1870. Brothers and sisters of the half blood were barred from inheriting as heirs. And, in the dictionary's flat phrasing, "a bastard can be heir to no one."

That last one is the important one, and it cuts both ways. An illegitimate child inherited nothing automatically, ever. So if an illegitimate child appears in an English will, the testator put them there deliberately and by name, which makes a will one of the better places to find an acknowledgement of a relationship that the parish register recorded only as a blank. The absence of such a child proves nothing. The presence of one is a decision.

This is worth holding next to what the same situation produced north of the border, where a kirk session would summon the mother and interrogate her about the father as a matter of church discipline. England produced no equivalent routine court record of paternity. What it produced instead, unevenly and only where a man chose to, is a line in a will.

One caveat, stated honestly. The dictionary is a secondary source, and we have not set these three rules against a second independent authority. Treat them as the shape of the law rather than as a citation you could rely on in an argument, and check the specific rule for your county and date before you build on it, since Kent's gavelkind and the borough English custom both departed from the general rules in ways the same dictionary flags.

Where there is no will there is often still a file

Most people did not leave a will. The National Archives notes that not everyone did, and that not all wills needed to be proved at all: a person leaving small amounts of goods could simply leave instructions with family or friends, and nothing reached a court.

Where there was no valid will, the court could grant letters of administration, appointing somebody to distribute the estate under fixed rules. For the Prerogative Court of Canterbury these grants are registered in the Administration Act Books, series PROB 6, covering 1559 to 1858. The National Archives is candid that grants of administration are "not very informative" compared with wills. They usually give you a name, a date, a place and an administrator, and the administrator is the genealogy: it was normally the widow or the next of kin, so the relationship is implied by the appointment.

The distribution rules themselves are worth knowing because they tell you what an intestate family's shares should have looked like. Under the Statute of Distributions of 1670, one third of the surplus went to the wife of the intestate and the residue was divided in equal portions among the children, with advancements brought into account as described above. Where there were no children, the wife took a moiety, meaning a half, and the residue went to the next of kin. Where there was no wife, the whole estate was divided equally among the children. The statute also makes a point of saying that the heir at law takes an equal part in that distribution even though he has separately taken land by descent.

There is one more document type that catches people out. A nuncupative will was an oral will, spoken rather than written, and the Statute of Frauds of 1677 hedged it about with conditions that are themselves evidence. Section eighteen provided that no nuncupative will was good where the estate exceeded thirty pounds unless it was proved by the oaths of three witnesses at least who were present at the making, and unless it was made in the time of the testator's last sickness, in the house of his or her habitation or dwelling. Section nineteen barred any testimony offered more than six months after the words were spoken, with one exception: where the testimony or its substance had been committed to writing within six days of the will being made.

Section twenty is the one to look for. No probate of a nuncupative will could pass a court's seal until at least fourteen days after the testator's death, and none could be received for proof "unlesse Processe have first issued to call in the Widow or next of kindred to the deceased to the end they may contest the same if they please." So the file should contain a citation summoning the widow or the next of kin by name, generated by statute, in a case where no written will existed to name them.

Taken together, a nuncupative will tells you without any inference at all that the testator was in his final illness, at home, that three named people were standing in the room, that the words were spoken within six months of being sworn to or written down within six days, and that the widow or next of kin was formally called in. That is a deathbed scene with a witness list, a date range and a named survivor. Few documents of the period are so specific about a moment.

What KleioBase does with a will, and what it does not

Disclosure, since we sell a tool.

A will is a single document that makes a dozen separate claims about different people, and the work is keeping those claims attached to the document that made them rather than flattening them into a tree. KleioBase is a research companion that turns scattered records into one connected archive: you upload the will image, its people, dates, relationships and roles are extracted, and the original wording is kept alongside so the bequest that names "my daughter Sarah, wife of Thomas Reade" stays readable as a sentence rather than becoming two disconnected profile fields. Before processing, the upload workspace has a Context box for notes about the document, its language, where it is from and its approximate year, which is where you say that this is a registered copy of a 1741 probate rather than a parish register. Every extraction goes through a Review step, where the fields are editable, before anything is committed to your knowledge base.

Once confirmed, the will becomes a record attached to each person it names, and where it disagrees with another record both claims stay visible against the document that made each one. That matters here more than usual, because a will's implied relationships and a baptism register's explicit ones frequently conflict. If you searched an archdeaconry's act books for the missing children and found nothing, the negative evidence feature records exactly what you searched and over what range, which is available on every plan including the free one, and it is written into a GEDCOM 7 export so it travels if you leave. That habit of writing down what disagreed and what came up empty is the same discipline behind the evidence layer your research log is probably missing.

Now the honest limits. Three of them.

KleioBase holds none of these images itself. The National Archives, county record offices and the commercial sites are where the documents live, and you have to obtain the page before you can upload it, which for a PCC will means paying somebody. Nothing in the software knows probate law: it will read a bequest to a witness on an 1840 will without knowing that section fifteen of the Wills Act voided it, and it will not flag an absent child as an advancement question. And the family tree has no structured way to record "named in the will but the relationship is inferred from a bequest," so a relationship you have reasoned to rather than read has to live as a note until you are satisfied it is settled. Deciding what a document proves is still the researcher's job, and that is the half of this work that does not automate.

Where to start with the will you already have

Take a will you have already read and go through it once more with four questions in hand.

Count the named children against every baptism you can find for the couple, and for each one missing, look for a settlement, a deed or a bond rather than a burial. Write down the three or four witnesses and find out who they were, because they are the one group the law guaranteed were not beneficiaries. Note which court proved it, and use that to work out where the property was rather than how rich the family was. And check the date of the will against the date of probate, and resist filling the gap between them with a guess at a death.

Then, if the will is a PROB 11 image, remember what you are actually holding. A clerk sat down with a dead man's paper and copied it into a register so that the court would have its own record. He was not preserving a family. He was closing a file. That the file happens to contain the only surviving statement your ancestor ever made about who mattered to him, and in what order, is a by-product of an administrative habit nobody involved thought would outlive them.