Some more handwriting and learning not to trust Ancestry transcriptions

[Another2Cats]

This is mostly a rant about Ancestry but I do have some specific questions about handwriting at the end.

I recently started a thread about handwriting here

https://www.mumsnet.com/talk/genealogy/5236815-anyone-here-good-at-deciphering-old-hand-writing-very-old-parish-register

and I just want to thank everybody that replied.
.

One suggestion, from @Shetlands, made a real difference to understanding things:

"Something I've always done is look at the rest of the pages and compare words so it's easier to decipher what they are. It can be tricky when the records are done by different vicars / curates but you can usually find enough done by the same hand."

I followed this suggestion and once you see how known words (eg common names or months of the year etc) are written then it makes it a lot easier to recognise how individual letters are written so you can then spell out other words.

So, doing this with the parish register that I was looking at I came across something a bit odd. I decided to check what I was reading in the parish register with what the record on Ancestry showed.

This is where I came across a huge problem.

On Ancestry I searched for any entries for this parish in a particular year and compared those entries with the scanned image of the parish register.

Lots and lots of the entries in the parish register were simply not recorded by Ancestry at all or were so different from the actual name (examples below) that you simply could not guess what the actual name was. There were also baptisms recorded as burials or marriages recorded as baptisms etc. There were entire years that had not been transcribed at all.

It appears that the transcriptions were done by different people. On one page, half of the entries might not have been transcribed and the other half were totally wrong. Then, on the next two facing pages the transcriptions were absolutely spot on in every detail (apart from a couple of minor mistakes) but then turn the page and it’s back to the rubbish again.

I have no idea if this is a one off or it has happened with other documents as well.

I checked very briefly some other nearby parishes. Some seemed spot on and some seemed rather patchy but nowhere was as bad as this. Maybe this was just a one off? I don’t know.

But it has made me seriously reconsider how I use the search function on Ancestry going forwards

Where I have reached dead ends or brick walls, in some cases it may be that Ancestry have either incorrectly or not transcribed the records at all. They may exist (like in the example of this parish register) and Ancestry have simply done a rubbish job of transcribing them.

It really has made me reevaluate what I thought I knew about Ancestry and has got me thinking that I need to now go back and double check all of the other early records that I casually assumed were accurate when I first started out on my search for my family tree.
.
.

Some examples of the poor transcriptions I came across

The baptism of “Somma Trifominge” was

“The christeninge of Thomas the sonne of Edward Viserd…”
.

Similarly, the burial of “Tho Heriftoning Gantoy” where the father was “Olive Gantoy” was actually

“The christeninge of alice hankox was in october the iiii day”
.

The wedding of “Tho Fomingo” and “Haboll Eddan” was actually

“The christeninge of Isabell the daughter of walter myles & margaret his wife was in december the xv day”
.

“Joen Maride” was buried and the father was “Edwards Maride”

The actual text was

“Wylia grevestocke clarke [he was in the church] was maried to Jone Edwards in november the xv th daye”
.

The burial of “Hynes Wilddonge” was actually the burial of “Agnes grevestock wyddowe”
.

The burial of “Wylia Postboye” was

“Wylia norris a pore boye was buried in Januarie the xi th day” [looking back at earlier records he was 13 when he died.]
.

So if you are looking for any of these families then you aren’t going to find these baptisms, marriages or burials on Ancestry if you just use their search function.

But maybe you will now? My DH insisted that we update ones that we found were wrong, so hopefully now people searching will be able to find them.

One that did make me smile though was the baptism of Joan to parents “Homer Wyfo” and “Elizabeth Wyfo”

In reality, the parents were:

“thomas bruer & Elizabeth his wyfe”

Whoever transcribed this thought that “wyfe” was the surname (but miscopied it anyway as Wyfo)!

I’m aware that Australians often use an “-o” on the end of words as a diminutive. When I saw this I just had visions of some one saying in an Australian accent “G’day, I’m Homer and this is Liz the wifo”.
.
.

@YellowPixie said

"Brilliant though that the priest/minister has recorded the name of the mother as far too often you just see "Elizabeth Smith daughter of John" and the mother doesn't even get a mention."

Yes this does make a huge difference, especially when the men have the same name, live in the same village and are having children at about the same time.

Sometimes the entries went into a lot of detail. For example, from a neighbouring parish I noticed this entry

“Willm Nicholas, the bastard sonne of Willm Nicholas of the pirsh of Barkeley yeoman, and Elizabeth Dirett his late servant, was baptized on the xxii th day of April”

The parish of Berkeley was about 12 miles away with its own church so I would guess they were trying to keep this from the neighbours.

But some entries just made me stop and consider how things have changed.

“The christeninge of Jone … was in September the v daye and was buried the same day”

I was particularly struck by these:

“The buriall of Margaret Morgan in childbed decesed was in September the xxi th daye”

then four years later

“Elizabeth Viserd decessed in childbed & was buried in Julye the xi th daye”

A few years on from that, a Richard was baptised on 1st June, his mother Agnes (aged 41) was then buried on 3rd June and he was later buried on 28th June

All three of these women’s deaths are in the parish register that is viewable on Ancestry but they have not been recorded so are not searchable on the website other than knowing to look at a particular page of the parish register.
.
.

Anyway, I did also say that I had some questions about handwriting; and I do.

The first is about a name that I see commonly written as Wilia or Wylia but the letter “a” has a long curved line that goes over the top of it. I’ve included an image below of an example of this.

Does a curved line after a vowel indicate a letter m or n follows it? So should Wylia be read as “Wilia” or is the curved line after the “a” indicating it should be read as “Wiliam”?

I noticed this most often with the name Wylia but I did also come across the same thing in a will from the same time and the word “Item”. Sometimes it was written as “Item” and sometimes as “Ite[curved line]”

I’ve included an image of this as well. The will appears to read.

"Item I give more to margret my daughter my best brasse pot"

"Ite[curly line] I give to Joane pit my daughter my lycke brase pan"
.
.

The next handwriting question is about Greek letters. I think a name might be meant to be “Christofer” but it is written as “xpofer” and the letters x and p have a line above them.

Could this be the curate using the Greek letters chi and rho (for Christ) and simply writing them in the same way that he writes the letters x and p?

I’ve attached another image and I believe it says:

“xpofer Balinger was married To agnes Stallerd noveber ix th”

Was the name Christofer ever written this way or am I reading too much into this?
.
.

Then finally, I came across four women whose names appear to be spelt as either Emmot or Emote.

Was this an actual name or am I just getting it totally wrong? Again, image attached. Was this a real name that has perhaps now just died out or am I missing something very obvious?

I’ve attached an image that I believe says

“The christeninge of thomas the sonne of Emote pitte unmaried was in december the xxviii th daye”

With the other women, their names were spelled Emmot.

Sorry this turned out so long, I just started typing and it all sort of just came out. I was just rather annoyed at all these missing records I came across. But any help anyone can give with the handwriting I would be very grateful for.

You can submit a report under 'More actions' when you're looking at the transcript. Make sure to scroll right down to the bottom and click on Submit when you've filled out the correction. Whether they'll do anything - they have made some of the corrections I've submitted, but my great-grandmother's occupation on the 1939 Register is still wrong.

Took me a while to find my maternal grandparents in 1921 because the surname had been wrongly transcribed, as had the eldest daughter's birthplace.

I haven't found my paternal grandfather and his mother in 1901. Found most of his siblings, who had left home by that time.

Parts of some censuses had gone missing long before the technology to copy them existed. I think 1911 was one. And of course 1931 was completely destroyed in a fire, so there'll be no more censuses until the 1951 is released.

Eg, if he lived at 210 then I can find records for 207, 209, 2011, 2013 etc but there are no records for 206, 208, 210, 212 etc.

The other side of the street may have been in a different enumeration district, in which case the forms would have been collected separately, and copied up in a different volume, which might have been lost, perhaps. Looking at the original records, have you scrolled back to where it describes the district, to see whether both sides of the street were included, and what route the enumerator would take when he walked round collecting the forms?

@Another2Cats I'm very interested in your reflections about transcription errors on Ancestry. I've never been able to find my grandmother on the 1939 Register on Ancestry, even though I know she was living in Wales at the time. Do you know if it's possible to search for a particular address? I don't have a current Ancestry subscription because I'm currently doing MyHeritage.

Thanks for your suggestions.

"...have you scrolled back to where it describes the district, to see whether both sides of the street were included, and what route the enumerator would take when he walked round collecting the forms?"

I'll have a look at that but you've prompted me to go back and check some other things as well. Thanks.

You can certainly search on Ancestry but I find that FindMyPast is easier to search for a particular address. I don't know if you can search for an address on MyHeritage.

If you want to send me a pm then I can have a quick look for whatever the address is to save you having to spend money on another subscription just for one thing.

Thank you, I'll PM you.

I have no idea if this is a one off or it has happened with other documents as well.

no it’s not a one off

i was working in an archive that went into a project with ancestry to get parish registers online and indexed

the parish registers were on microfilm in Salt Lake City and they were put on reels, so you’d have several parish registers all on one reel. For example 10 registers and started with Appledore, Bideford, then Barnstable and so on until the reel was full and they started another at Landkey

so whoever transcribed this register in India, wouldn’t understand that the entire film wasn’t Appledore, but put all the transcribed records down for all 10/12 parishes as Appledore.

then they couldn’t understand the sentence just maybe letters and words

so we had Nicolas Bastard son baptised son of Maria Turner so unless you searched for Nicolas Bastard in Appledore

the entire county was transcribed mess

the ciunty used family history volunteers to try and make corrections

it wasn’t Devon where I worked or this happened- I’ve used it as an example

the archive were devastated at the mess and have never done another project with them.

If you search for Base or Bastard as a last name on Ancestry dot com, you get results in the parish registers

incidently ive been using ancestry over Xmas for census and the transcripts for place where incorrect

I had on the original census street, Spon End city Coventry

byt the census transcript told me it was the parish church off St John Leamington Spa

MissScarletInTheBallroom

1939 register was used by NHS up till 1991 so if the person was still alive until then some names will be retracted until their birth has passed 100 years

therefire if your grandmother wasn’t born until 1925 or after she will not or shouldn’t appear unless they died before 1991

1939 also puts women’s future names, as they married and NHS records were updated they wrote in the future name of the person

An archive office that I knew was in the County Hall. There was a prison behind it. Supposedly some early indexing work, to create a card index of some records - long before the internet - was done by some of the prisoners. Of course those errors will be replicated if the card index is copied.

Some errors do go back to the original census process. The forms filled out by householders were transcribed into volumes, and for the earlier censuses it's the volumes that were preserved and later microfilmed - the household schedules weren't kept once they'd been transcribed. There was room for error in the transcribing process. One I've seen more than once is where the transcriber 'dittoed' the birthplace too many times.

Something else that happened is where the householder was illiterate and also not a native of the place s/he was living. The enumerator filled out the form on his/her behalf, trying to make sense of unfamiliar names and birthplaces spoken in an unfamiliar accent.

'Lambrus' and 'Tisus' are two birthplaces I've seen in the original records. That is the local pronunciation, but the enumerator, in a different town, presumably wasn't familiar with the places. Local people will know what was meant, but people who don't know the region won't.

(At the time of at least one census, older schoolchildren were taught how to fill in the form, in case their parents were illiterate)

Thank you.

I would understand my grandmother's record being closed (although my other grandmother and her sister's records were both open even though they were younger and died more recently). But my great grandmother doesn't appear either, and she died in 1984.

"...the County Hall. There was a prison behind it."

Totally off-topic, but that sounds like Oxford? (Not expecting a response, by the way - I just thought that I remembered reading that somewhere)

No, not Oxford. Though looking it up, I see that Oxford prison was close to County Hall there.

(Oxford prison is now a hotel which IIRC featured in an episde of Lewis.)

I found my grandfather on the 1939 register, eventually

his name was Illam instead of William
sorry if I’m teaching my grandmother to suck eggs

but put no names in the search index and use dates of birth, places instead

Thank you, I will try that.

Ignoring the people being rude about genealogy...

Ancestry is (in)famous in the genealogy community for rubbish transcription. This is because they have discontinued their human transcription project and are now doing it all by OCR - computer programmes trained to "read" the documents and transcribe. SOmetimes they are fine, sometimes they output complete nonsense. One of my favourites is Sarah mistransribed as Laser. It makes searching very difficult, even using wildcards.

FamilySearch on the other hand sticks to human transcription (you can join in by hitting the "get involved" tab on their homepage). I am both an indexer and checker for them. Everything is transcribed, and checked by a separate person so there are at least 2 sets of eyes over each record. And you don't get to be a checker until you've done a LOT of indexing.

Another prime example attached - Ancestry does allow users to flag errors which someone has done here, clarifying that "Willand Testament" is clearly not someone's name. But they never correct it completely.

As a family, we helped to transcribe the 1881 census, it was done through our church. As a kid at the time, older teenager, I have no doubt I got some of the spelling wrong, and would say thr same for everyone, we tried our best, but reading what was a photocopy of a transcript was hard

Projects like FreeCEN and FreeBMD use double transcription plus checking, if you ever use the index searches there you can click on the magnifying glass and see who transcribed it.

I've found FreeBMD fairly good on the whole, and the one time I reported a transcription error they corrected it.

Wouldn't trust Ancestry all that much.

I know in the past Ancestry did have prisoners transcribing their records so not sure how motivated they would have been for accuracy!

(This is totally off-topic - apologies)

That's interesting, I wasn't aware of that.

"...so not sure how motivated they would have been for accuracy!"

It all depends. I've worked with a number of male offenders over the years. You get all different types.

At one end of the scale there are people that I have had to cover really basic maths with. For example (and this is a real example that has happened more than once):

Very basic things like fractions can be a real problem for some people. I have literally had to explain why 1/4 is a smaller amount than 1/3 even though the number "4" is "bigger" than the number "3".

Yes, some people really do think that 1/4 is bigger than 1/3 (because 4 is bigger than 3).

I ended up by demonstrating with 12 pieces of wood (this was during a woodwork session for the prisoners) that if you divided them up into four equal groups then you had fewer pieces of wood in each group than if you divided them up into three equal groups.

At the other end of the scale, especially among older prisoners, you can get quite a lot of very well educated people.

These prisoners will see something like becoming a library orderly (basically, a library assistant) as a real prize job, a lot better than perhaps being a wing cleaner or even working in the prison kitchens (another prized job due to the higher levels of pay), or learning bricklaying or some other practical skill.

Other similar jobs that prisoners like this will look for are orderly jobs in education.

I can well imagine that if a transcription job were given to prisoners like this then they would likely welcome the opportunity (and probably do a much better job than a non-native speaker).

However, expecting HMPS to organise this in any sort of effective manner is probably expecting too much. If Ancestry or any other organisation wanted to do this then it would probably be better if they approached a specific prison rather than relying on HMPS.

/End Rant

I've never heard of transcriptions being done by prisoners, tbh.

Is the current human transcription done by transcribers whose first language matches that of the records?
From what I've seen in archive transcription errors the period of human transcription was particularly bad but not surprising if the don't understand the language let alone ancient variations. Its hard enough when you do speak the language (IME).

My personal hell for trying to interpret records for one branch of the family has been the Drouin collection - tons of errors and large blocks unreadable in the online versions and so impossible to consider the quality of the transcription. The large number of place name changes over the past couple of centuries and the use of old Quebecois doesn't help either 😅

As far as I know it’s now all OCR. The archives project was human transcription and closed a couple of years ago. You are allowed to transcribe in your non native language - I often do Spanish records for family search as here is always a huge backlog