Tip When Checking Old (Late 1700s and Early 1800s) Sources:

When looking at alphabetized indexes, don't assume all the entries are in strict alphabetical order. Indexes that purport to be alphabetized generally alphabetize things accurately by the first letter, but within each letter the entries are sometimes out of order. I just noticed this in several indexes from that era, and thought I'd pass it along.

Anonymo the Anonymous:
I'm curious why. The alphabet was in the same order back then, right? And since you're singling out records from the 1700s and early 1800s, it seems this isn't an issue common to all eras before computerized records. Was it just not customary to care about alphabetizing records beyond the initial character then or what?
7.14.2008 4:52pm
Stephen C. Carlson (www):
It's not limited to the 1700s and early 1800s. Prior to that, it was common for indexes and even dictionaries to be partially alphabetized.
7.14.2008 5:03pm
Yes, genealogists are familiar with indexes done in this fashion. Sometimes one finds other methods used, e.g., alphabetization by first letter and first vowel.
7.14.2008 5:07pm
PatAtty (mail):
Is the issue related to U and V? I've noticed that the alphabetical ordering of T, U and V was not consistent during the late 1700s and early 1800s.

See Johnson's 1824 Dictionary
7.14.2008 5:34pm
andy (mail) (www):
Weird...couldn't they just use the auto-alphabetize function in Word?
7.14.2008 5:55pm
kevin r:
Not only U and V, but I and J were often combined. Noah Webster was one of the first to really treat I and J as separate letters, and not everyone immediately took to the idea. You'll notice that 1824 dictionary doesn't have "J" in the alphabet. It might have written the character "J" in words like "justice", but would have considered it just a variant way to write "I" when being used as a consonant.
7.14.2008 6:00pm
Michael F. Martin (mail) (www):
Evolution of language away from latin.
7.14.2008 6:24pm
James Lindgren (mail):
Handwritten probate records are the same way.

When the courts had a contemporaneous index, they usually had a book with one page for each letter of the alphabet. When they got a new case, they would add that case to the correct page of index book.

So the cases were usually in rough chronological order WITHIN each letter of the alphabet.

I'll bet that the indices you note were generated by a similar process of adding to a list alphabetized by the first letter of the last name or topic.

Jim Lindgren
7.14.2008 6:48pm
Eugene Volokh (www):
It wasn't just I/J and U/V.
7.14.2008 7:08pm
Approximate alphabetic order? Yeah, just like in the hard copy files at the law firm I left...
7.14.2008 7:34pm
The Ace:
In my worst summer job ever (the summer before law school), I was a title searcher in a jurisdiction lacking computer capability to search title history.

When titles were indexed prior to he age of computers, (pre-1980) they were CATEGORIZED alphabetically but filed chronologically. Thus, if Eugene Volokh conveyed an interest in real property, the transaction would be indexed in a book entitled "VO" but then would be written in the index ledger in the next open space. If the next day George Voinovich sold an interest in real property, his name would be listed after Eugene Volokh's although it should be before it if it were alphabetized. I have no doubt something similar to this system operated in other indices.
7.14.2008 9:26pm
Michael F. Martin (mail) (www):
The real question is whether there is a better way to organize words than alphabetical order.

Actually, google sort of answered that question, didn't it? The best way is to be able to type them into a box and hit a button to get the definition. three hundred years from now somebody may be writing about how strange it is to see how people who lived in the late last millenium liked to sort their lists of words into a mostly stable order beginning with A, B, C...
7.14.2008 9:35pm
John (mail):
OK. We've seen all the old quotes. What are you up to rummaging around the nineteenth century?
7.14.2008 9:41pm
MarkField (mail):

I'm curious why.

Basically because the system of paper and pen was too cumbersome. Imagine a bound book with a few hundred pages. The county clerk has to enter all relevant events in it (court minutes, say, or marriages or deeds or taxes paid). He (always he in those days) had no way to anticipate how many transactions there would be under each name. What he did was make an educated guess about the relative frequency of the first letter of surnames, and then enter them using just the first letter rather than full alphabetical.

Why not go back and insert new names between the old? Well, the books were bound (storage systems then and now are not friendly to loose leaf) and binding was expensive. He couldn't just add in new pages as needed. Moreover, suppose he ran out of room for a name such as "Fairfax" (easy to do in Northern VA). Since he couldn't insert new pages, he'd have to re-enter all the previous transactions after inserting the new one in order to preserve the pure alphabetical order. Too time-consuming and expensive.

Those of us who do genealogical research become all too aware of issues like this. I personally am dreaming of the day when all the old records are scanned in searchable text and a few queries will give me all the information I want....
7.14.2008 9:43pm
Tom Tildrum:
There was also the "s" that looked like "f". That alwayf consufed me.
7.15.2008 12:25am
The s that looked like f only occurred in the middle of words. So this is correct, and thefe also, but thif ifn't.
7.15.2008 1:10am
Bill Poser (mail) (www):
I imagine that this was because alphabetizing a list of any size by hand is extremely tedious and time-consuming if you subsort on every letter. A sort just on the first letter can be done using the bucket sort technique, in which each item is put into one of finitely many bins. While comparison sorts can be no more efficient than O(n log n), a bucket sort has complexity O(n). Note furthermore that when writing material was valuable, it was desirable to minimize scratch paper or slips for each entry. A bucket sort also minimizes writing material.
7.15.2008 1:37am
David E. Young (mail) (www):
I have some experience with the type of index probably being discussed here.

I suspect the type of index Prof. Volokh refers to is like that in Elliot's Debates. There is a relatively simple reason why such an index should not be completely in alphabetical order compared to a modern book index. Most modern book indexes tend to be very simplified compared to those from these earlier periods. Often only words and very short phrases are indexed in many modern books. Thus, it is much easier to list everything in the index alphabetically, and modern readers are often more used to such indexes.

In his Debates, Elliot indexed statements of occurrences, facts, and ideas as incomplete sentences. There is no good reason to list any of these alphabetically. The document collection is laid out in chronological order, so the index entries under the alphabetized headings are often presented chronologically - the order in which they happened. This results in the entries under any particular heading having constantly increasing page numbers as one goes through that heading's entries.

This older type of index is also found in some modern historical document collections. One example would be the index containing volumes of The Documentary History of the Ratification of the Constitution from the Wisconsin State Historical Society. Another example would be my collection, The Origin of the Second Amendment.

I also used this type of index for my history of the Second Amendment, The Founders' View of the Right to Bear Arms. In dealing with a historical subject presented in chronological order, it would not be beneficial to present entries that are fairly complex and have highly random wordings in alphabetical order. It is much more helpful, especially in dealing with historical subjects, to present the entries in chronological order.
7.15.2008 10:58pm