Making books in Low-Tech & LT:Companions 1 & 3

[Jussi Kenkkilä]

I was interested in calculating some stats for low-tech libraries, but ran into problems with the numbers. To double-check them, I grabbed a few examples of manuscripts and data on book production.

From a dataset of about a hundred manuscripts gathered from various sources (mostly codicology books online) I calculated a few reference values. (The dataset included books of different sizes, mostly quartos and octavos, written in Europe, North Africa and Middle-East in various languages.)

The average page count was 560 (230 folios) and was pretty close to this value for 4s, 8s and 16s sizes. The average words/page was 90 (for quartos in the set 140, and for 16s & 32s smaller). The average words/manuscript was then about 50 000.

Comparing these to the values in LT, HT and LT:C1 (500 words/page printed, 125 words/page handwritten) seemed close enough (several sources put modern words/page average in the 400-450 range and modern book lengths in the 70-100k words range).

Back-calculating from the weight of paper I arrived to the result that a "small collection" contains 20k hand-written pages. Therefore it consists of ten 2000 page manuscripts, probably split into four 500 page books.

So far the numbers match nicely. However when calculating the price from the time it takes to produce the books (mostly the scribe-work), the total costs for low-tech libraries and books seem too low.

From various sources I checked, the speed of scribing was in the range of 1.5-10 folios/day, mostly quoted as 2-3 folios/day. With an average speed of 5 pages (2.5 folios) per day, it would take 50 days (2 months) to scribe an average (500 page) book.

As several sources (and at least one primary source) claimed that the scribe-work accounted for half of the final cost, I used the 0.55 ratio of labor to price to calculate the labor costs. From this I get $770 to scribe a single book and $1400 for the finished product.

Since the manuscripts used as reference were written on parchment, I checked the total cost of the parchment used (250 leaves). It was $150, which is about 1/5th of the scribing cost (and the value of materials calculated from labor).

A small collection of manuscripts would therefore cost $1400*4*10=$56k, a lot more than the $3500 estimated in LT:C1. The parchment alone would cost $6000.

Since most of the higher cost comes from scribing, changing the media to paper ($600) wouldn't help.

To check my numbers, I re-read everything in the relevant chapters and realized that the assumed writing speed was 4000 words/day. Calculating the scribing rate from the previous sources, gave me significantly lower rate: 1250 words/day (if scribing 10 pages/day at 125 words/page). To verify this, I checked some new sources on scribal speeds. Those mentioned writing cursive at a rate of 8-20 folios/day (average 28 pages/day) and a dataset of samaritan scribes (in "semi-cursive" script) averaged at 10 pages/day (confirming the numbers given by several sources). From this I noted that the 4000 wpd rate would probably be achieved with cursive style.

Still even at the rate of 4000 wpd, it would take 16 days to scribe a book, at a labor cost of $240 and a total cost of $480. It is still high compared to the cost of parchment, but would result in the "small collection" valuing close to $20k.

From these I would suggest that the real cost of a low-tech (scribed on parchment) is multiplied by 16 (like very fine quality, 4 for writing speed & 4 for lower words/page count to achive the same information content) in addition to the 10 in LT:C1. To match the prices of writing materials, they should be multiplied by 5.

For printing on paper at a rate of 250 pages/day, the price of a book would be $137 ($62 labor + $75 paper, close to the estimated ½ of total costs for early printing). A "small collection" would be about $1400, which is about 4 times the High-Tech cost. Increasing the printing speed to 1000 pages/day and dropping the paper price multiplier gets us close to the HT price. We could say that the price drop on paper comes from switching from rag to wood pulp paper.

Comments? How do my assumptions differ from the authors? Do the differences come from using differents sources for the numbers?

I ignored the cost of binding and covers, although I calculated some costs to make covers (ca. $31 for ½" oak covers with leather on top, including labor, around 2% of total), and renting an exemplar (6% of the total costs).

[Varyon]

First off, the library costs assume printing - multiply by 10 for handwritten manuscripts. That's $35,000, and 1000 lb, for the basic library. Assuming negligible weight for binding material, that's 100,000 sheets and $3,000 in paper. At 125 words per sheet, that's a whopping 12,500,000 words. A professional scribe is going to have effective skill around 18, maybe 20 for a really good scribing shop (DX-based scribe of 12, +6 to be legible, up to +2 for the shop) when writing at a rate of 4,000 words per day. That's going to take 3125 man-days, or 125 man-months (over a man-decade). The scribe could probably take -6 for -60% to time necessary, for 50 man-months. LTC1 notes that scribing is almost the entire cost for such documents, so you're looking at $40,000, which is pretty darn close to the $35,000 estimate. Your estimate of 20,000 pages for 250 lb would put us at 80,000 pages and 10,000,000 words. Taking the -6, the scribes would be able to churn something like that out in 40 man-months, costing $32,000 - adding in the cost of paper, you get to exactly $35,000.

[ArchonShiva]

I love GURPS.

This really reinforces why religious books were so prevalent compared to non-religious books, though: the church didn't need to pay its monks the going rate.

[dcarson]

The other source of free labor is students renting a book for a course and copying it themselves. That was a standard way of students building their own library.

[Peter Knutsen]

Originally Posted by ArchonShiva

I love GURPS.

This really reinforces why religious books were so prevalent compared to non-religious books, though: the church didn't need to pay its monks the going rate.

The interesting question here is how rare scribal ability is.

Out in the mundane medieval world, it was quite rare, and so scribal work was expensive. Very few townsfolk were literate and able to do decent caligraphy, and so they could demand high wages, well above the cost of materials, mainly parchment and ink.

But in the monasteries, you had a significant minority of monks (a majority in some monasteries, or at least close to it) who could perform this work, and all you had to do was to feed, clothe and roof them, and cheaply at that, since they were supposed to live austere lives. They didn't always, but even in luxury monasteries, their upkeep still wasn't very expensive.

Also, to the OP: Are you familiar with an early online work by S.John Ross, titled GURPS Books or GURPS Libraries or something? I think it is still somewhere on his The Blue Room web page, or if not then you may have to try to find it via the Wayback Machine.

[Jussi Kenkkilä]

Originally Posted by Peter Knutsen

Out in the mundane medieval world, it was quite rare, and so scribal work was expensive. Very few townsfolk were literate and able to do decent caligraphy, and so they could demand high wages, well above the cost of materials, mainly parchment and ink.

Also, to the OP: Are you familiar with an early online work by S.John Ross, titled GURPS Books or GURPS Libraries or something? I think it is still somewhere on his The Blue Room web page, or if not then you may have to try to find it via the Wayback Machine.

The literacy rate does affect the prices of books, but the high price of writing materials will keep them high even if you could set up a work-house of literate poor or slaves to do the scribe work. As there were various shorthand writing styles in the antiquity, the cost of scribing could be pushed very low at the price of legibility.

The "Book Rules for GURPS" is still online. I'll check it and apply them to the current edition.

[Polydamas]

Originally Posted by Jussi Kenkkilä

The average page count was 560 (230 folios) and was pretty close to this value for 4s, 8s and 16s sizes. The average words/page was 90 (for quartos in the set 140, and for 16s & 32s smaller). The average words/manuscript was then about 50 000.

I studied some medieval book history back in my undergraduate days, and 230 folios is pretty high for a medieval book. There definitely were things like "the Old Testament in one volume" or illuminated chronicles of 216 folios, but it was much more common to have one volume of the psalms, one volume of the gospels, and a notebook with your family's secret recipes and words of wisdom. You might want to check that your source is not mixing up printed books and manuscripts. Ninety words per page sounds like it might include manuscripts which had more picture than text, or very small manuscripts. Fifty thousand words per volume seems about right.

Rosamond McKitterick has some estimates for early medieval prices using Diocletian's Edict of Maximum Prices in The Carolingians and the Written Word. They are pretty hand-wavy but would be fine for gaming.

[Jussi Kenkkilä]

Originally Posted by Polydamas

I studied some medieval book history back in my undergraduate days, and 230 folios is pretty high for a medieval book.

I haven't ran into any systematic study of manuscript lengths, but I made a (semi-random, ie. first books of the alphabetical pages) sampling of the Parker Library (http://parkerweb.stanford.edu/parker/) online manuscript data for comparison. The average of this small sample was about 160 ff, or 320 pages. Since the variation on the whole dataset is very high (standard deviation of 170, min. 13, max. 944), I too am sceptical of the average page number. If anyone knows of a good study on the subject, please let me know. Otherwise I'll improve the data by adding more manuscripts from different collections.

Some manuscripts are bound into several smaller volumes, so we may assume that one modern book is equivalent to 4 manuscript volumes on paper, 8 (of ½ page count) on papyrus or vellum and 20 (of 1/5 page count) on parchment. This way the approximate thickness of each book/volume remains the same as the thickness is probably based on keeping the books of manageable size.

With this lack of data word count should be a better estimate for the information content (which should be the same for all libraries of the same size).

[Turhan's Bey Company]

Alas, the article I wrote for web-Pyramid on information density is no longer available, though the data-to-oxcart converter is still out there.

[Jussi Kenkkilä]

Originally Posted by Turhan's Bey Company

Alas, the article I wrote for web-Pyramid on information density is no longer available, though the data-to-oxcart converter is still out there.

Hi!

What values did you place your calculations on? (words/page, pages/book, pages/pound..)

Many of the library manuscript collections list the number of folios and even number of lines/page, but not how many characters per line.