How often were medieval scribal abbreviations used?

Question

In the preface to The elements of abbreviation in medieval Latin paleography, Adriano Cappelli writes

Take a foreign language, write it in an unfamiliar script, abbreviating every third word, and you have the compound puzzle that is the medieval Latin manuscript.

I would assume that this is just a joke, but after some reading of the wide use of scribal abbreviations (sigla, including truncations and contractions) elsewhere, I've had second thoughts. While it would clearly be impossible to pin down an exact percentage, is Cappelli's joke close enough? Were a substantial amount of words in medieval Latin abbreviated, or only a relatively small amount?

I understand from Scrittura e Civilta that the use and type of scribal abbreviations varied throughout the Middle Ages to such an extent that they can be used to approximately date manuscripts, but I do not think that this should impact the question too much.

Answer 1

Yes, Cappelli is basically right (which is perhaps not surprising, he being an authoritative figure in the field of palaeography).

Every third word is probably above average, but in most documents one seldom finds a sentence without any abbreviations. Parchment was not cheap, nor was scribal labour. The same applies to manuscripts in any European language, at least in my experience (mostly Latin and Dutch). Some genres are more likely to contain fewer abbreviations than others, such as literary works written in book script; but even an illustrated Bible will usually be full of them. A few arbitrary examples, with abbreviations in bold:

http://www.sothebys.com/content/dam/stb/lots/L12/L12242/009L12242_6CB6T.jpg

Regnante domino nostro ihsu xpisto cum post passionem suam resurrectionem et ascensionem eius conuersis ad dominum cunctis finibus terre nec in angulo occeani aliquis se a colore fidei absconderet in conveniendo populos in unum et reges ut servirent domino. fuit in britanni(a)e partibus rex quidam deo notus tam uita quam nomine qui religiosus...

That's 32% of words.

http://www.villa.culture.fr/mediatheque/communs/images/grandFormat/0/21_2.jpg

Pars est prima prudencie: ipsam cui precepturus sis estimare personam. Neque enim formator agricole debet artibus et eloquencia rethores emulari quod a plerisque factum est qui dum diserte loquntur rusticis assecuti sunt, ut eorum doctrina non a dis?tissimis possit intelligi. Sed nos recidamus prefationis moram: ne quos reprehendimus imitemur. ...

That makes 43%.

On the other hand, there are also manuscripts that use few abbreviations:

http://www.europeanaregia.eu/sites/www.europeanaregia.eu/files/manuscripts/regiacarol_lat_5763.jpg

Gallia est omnis diuisa in partes tres, quarum unam incolunt belgae, aliam aquitani, tertiam qui ipsorum lingua celtae, Nostra galli appellantur. Hi omnes lingua institutis legibus inter se differunt. Gallos ab aquitanis garunna flumen a belgis matrona et sequana diuidit. Horum omnium fortissimi sunt belgae propter ea quod a cultu atque humanitate provinciae longissime absunt minimeque ad eos mercatores saepe commeant atque ea quae ad effeminandos animos pertinent inportant proximique sunt germanis qui trans renum incolunt quibuscum continenter bellum gerunt ...

That's 11%.

The overall ratio is difficult to establish (and perhaps less necessary), but you can get a fair impression by looking at a couple of documents in Google Images.

Answer 2

The proportion of abbreviated words in a medieval manuscript depends on the time when it was written and also the individual scribe, but could generally be quite high. I believe that the estimate by Cappelli cited in the question is rather conservative.

As an example, consider the following excerpt from an 11th century manuscript, containing verses 137 to 143 from the first book of Ovid’s Metamorphoses. (It is the top of the first undamaged column; I didn’t compare the number of abbreviations to other parts before choosing this.)

(Pal. lat. 1699, page 1 verso, top of the right column; courtesy of Heidelberg university library)

Here is my transcription, with those words that contain abbreviations in italics:

nec tantum segetes alimenta-que debita diues
poscebatur humus sed itum est in uiscera terrae
quas-que recondiderat stigiis-que admouerat unbris
effodiuntur opes irritamenta malorum
iam-que nocens ferrum ferro-que nocencius aurum
prodierat prodit bellum quod pugnat utroque
sanguinea-que manu crepitantia concutit arma

If you count each enclitic -que (except as part of uterque) as a separate word, there are 46 words in this section, with 23 of them (50%) having been abbreviated. If you consider alimentaque etc. as a single word, there are 40 words, of which 22 (55%) contain an abbreviation.

In my (limited) experience, this is not abnormally high.