Image credit: see note 1

Remember my “Dear Santa” wish list post from last month? Well, Santa came, in the form of a cheery email with exciting historical source news from Clark House supporter, eagle-eyed historian, and all-around great friend Ellen Hickman.²

Ellen’s email, I’m happy to report, has opened new pathways to locate all kinds of otherwise hard-to-search primary and secondary historical records and documents from the Clarks’ era. Since receiving it, I’ve been chasing after documents down one formerly-obscure research “rabbit hole” after another, with some gratifying results…

New horizons in records access

In her email, Ellen asked if I knew that FamilySearch.org—the massive, free, online repository and access point for the millions of microfilmed historical and genealogical records created and archived by the Church of Jesus Christ of Latter-day Saints—was beta-testing a new Full-Text Search portal ³ to access their records? Well, no, I did not know about the new Full-Text Search. But I do now!

What is Full-Text OCR?

OCR, or Optical Character Recognition is the rapidly-improving technology that allows computers to, scan, read, and decipher printed, typed, or even hand-written text. And now, after the OCR process makes a machine-readable version of a document, the full text of that document (and not just a few keywords) is indexed and becomes searchable. FYI, I did a post on OCR back in August, 2024, called Bad Handwriting: not just a problem from the past. Click the link to learn more about the challenges of running OCR on the wide variety of print and cursive handwriting found in modern—and historical—sources (and be sure to enjoy the linked Tom Scott video, too).

So what’s up at FamilySearch? As Ellen surmised, it “looks to me like they have just OCR’d huge swathes of county and other regional records on microfilms. The search feature is very crude, it’s not at all clear what is in this collection, and you can’t seem to easily browse your way to specific records for a location. But the access to such a huge body of records at home is unbelievable.” For what it’s worth, that agrees with my initial experiences, too.

Here’s an example

For all its current shortcomings, the new full-text search can provide links to many potentially relevant documents. Let’s look at an excerpt from one document that I just found through the new search portal, an official county record of a deed to land that Jonathan and Mary Clark sold to one Wendell Hoffman in October, 1841. Here’s the beginning of the microfilmed text of that deed, not previously indexed, as found this week via the full-text portal on FamilySearch⁴:

And below is the computer-generated transcription of the same bit of manuscript text, as recently read and made searchable by the new, full-text OCR process:

! 4 This Indenture made the thirtyth ( 30 day of Oct ober in the year of our Lord one thousand Eighthe wit & forty one between Jonathan St Clark and Mary his wife of washing to Co Wisconsin Territory parties of the first part . and Wendell Hoffman of Sheboyga County Wisconsin Territory party of the second part […]

Far from perfect, but it’s a start. For those of you that don’t regularly read 19th-century deeds and mortgages, a reasonably accurate transcription of this bit would read:

[page] 213
This Indenture [i.e., deed] made the thirtyeth (30) day of Oct-ober in the year of our Lord one Thousand Eight hundred & forty one between Jonathan M Clark and Mary his wife of Washingto[n] Co Wisconsin territory parties of the first part. and Wendell Hoffman of Sheboygan County Wisconsin Territory party of the Second part […]

So, yes, the current OCR is pretty rough. But it found enough likely search words, like Jonathan Clark and Mary and forty one that, combined with other OCR-recognized search parameters (e.g., Sheboygan county, a specified date-range from the mid-1830s to about 1860, and keywords such as Deed or Mortgage), allowed the document to be located and added to my list of results. Which is great, as this deed—and others like it—opens a whole new vista on JMC’s life after Fort Howard, and his apparent use of land as an investment tool. I’ll have more to say about all this as I transcribe and assess these previously-unknown deeds and mortgages, created by JMC over the span of almost 20 years.

My impressions so far…

Yes, the current full-text search engine is clunky, and the OCR makes a lot of errors. But the new full-text search is, all the same, a really remarkable technical achievement. It manages to decipher enough text correctly (including scrawled signatures and other messy handwritten text), that it now makes a lot of material, such as otherwise unindexed early county deeds and mortgages, much easier to search, locate, and study. 

It has also led me to some neat Jonathan Clark discoveries among huge collections like the microfilmed Territorial Papers of the U.S – Wisconsin 1836-1848, one of many unindexed online microfilm series that have been too daunting for me to manually search through previously. And I’ve already learned a lot more about JMC’s activities during the “mystery years” between his leaving the army in October, 1836, and purchasing his first Mequon land in December, 1839. I’ll be blogging about this shortly, once I tidy up the OCR translations, and assess the contents, of a whole pile of handwritten 19th-century deeds and mortgages.

And that’s not all!

For what it’s worth, the new FamilySearch full-text search already seems to have OCR’d a lot of records for early Wisconsin and—I think—Illinois. New Hampshire seems to have quite a few full-text records too (lots of possible Clark ancestor leads, especially in Grafton County, NH) and maybe some good Vermont records (although the wilderness of northern Vermont always seems to lack much in the way of records until the 1820s or so. Perhaps this is due to the paucity of early 1800s settlers and record-making bureaucrats, or that the new search portal has yet to run OCR on old VT records. We’ll see.) I’ve also had some luck locating and reading various Turck and Gay family documents from New York, circa 1780-1840 or so (mostly wills and, especially, land records).

I’ve been dabbling with the Lower Canada records with some success, although a lot of the hits I’ve had so far are from the Land Petitions of Lower Canada which have already become more searchable (names and some keywords, only; not full-text) via the improved portal at Library and Archives Canada, and which I’ve been slowly sorting through on this blog for almost a decade. But I’ve only dipped my toe into these Family Search Canadian full-text OCR records, so we’ll see.

And along with local Clark, Bonniwell, and Turck records, I’ve been finding documents that should help solve some of the mysteries about other early Mequon settlers, including very early settler Isham Day, first postmaster John Weston (aka John Western) and Bonniwell Gold Rush expedition member Peter Rattery (or Rattrey)

One drawback (at the moment)…

Historians like to cite our sources. Proper citations allow us to credit our sources and easily locate and re-assess them as needed. And it allows readers to check our work, and make their own assessments of the information presented and its accuracy. The documents at FamilySearch, originally on microfilm, and now available digitally, have been a staple of genealogical research and citation for decades. In fact, a vast body of citation practices and familiar source names are based on the old FamilySearch microfilm naming and numbering system.

And that’s why it’s unfortunate that the new Full-Text OCR portal has devised a new, and more “digital” OCR-generated citation system. As an example, for the item above, the new Full-Search citation reads:

“Sheboygan, Wisconsin, United States records,” images, FamilySearch (https://www.familysearch.org/ark:/61903/3:1:3Q9M-CSV4-H9VT?view=fullText : Jan 10, 2025), image 110 of 126; .

On the one hand, this is an improvement, as up to now, this series of microfilmed deed books had no index at all. But I much prefer to know also where and when in the “Sheboygan, Wisconsin, United States records” this document originated. Fortunately, if you’re willing to scroll around a bit, you can work your way back to the initial “target” at the start of the digitized microfilm and discover that this document is from what used to be described as:

FamilySearch, United States, Wisconsin, Sheboygan [county], Deeds, Film #1392877, DGS #8117690, Item 6, Deeds vol. A, (1839-1842), pages 213-214.

This minimal⁵ “old-school” citation tells the reader both how to access the digital version of the document and, importantly, the original source of the document. I hope FamilySearch will improve its current, overly-broad OCR-generated citations in future iterations of its important new Full-Text Search engine.

Phew!

That’s all for now. I’ve already accumulated over 100 “new to me” early JMC- and Mequon-related documents via this new search feature, and I’ve got some transcribing and analysis to do.

My thanks again to Ellen Hickman, and I’ll be back soon with more Clark House History. See you then!

_____________________________

NOTES:

1. Image credit: Santa Claus sugar plums—U.S. Confection Co., N.Y. United States, ca. 1868. Library of Congress.
   
   
2. Full disclosure: Ellen is a direct descendant of Jonathan & Mary Clark and thus, my 4th cousin, once-removed!
   
   
3. To access any of the records at FamilySearch, whether via the new Full-Text Search portal, or using FamilySearch’s original search page, you’ll need to create a free user account and password. If you are interested in local history or genealogy, you’ll really want to create your free account and start searching through the available records. Most of their online records can be viewed at home. But some records, for contractual and/or copyright reasons, may only be viewable at a local Church of Jesus Christ of Latter-day Saints “Family Search Center” (formerly known as “Family History Centers”). Here’s a link to find a Family Search Center near you.
   
   
4. link to document: https://www.familysearch.org/ark:/61903/3:1:3Q9M-CSV4-H9VT?view=fullText
   
   
5. Exemplary, modern, full citations for items like this can be really, really long and complicated, and I rarely spend the time to make such exemplary, modern, full citations. But I do like to make citations that will at least allow me, or any researcher, to go back to the same document and assess its contents for themselves.