Salvete!
Can someone take a stab at ballparking the approximate frequency of the the different noun declensions? Just out of curiosity.
Thanks.
Hi, do you mean the percentages of the different declined forms (accusative, dative etc.)? If so, there’s a chart (12.1) in the Oxford Latin Syntax vol. 1 (which is full of all sorts of frequency data, mostly for verbs though because it’s a syntax, not a grammar). The chart is online, on the companion website to the syntax. The syntax says it’s ‘based on an analysis of c.6000 nouns (including pronouns) and noun phrases and prepositional phrases in a corpus of various texts’ (p. 1180):
If however you mean the percentages of words that are first declension, second declension etc., not sure, but someone should be able to extract this data I’m guessing from a comprehensive word list, based on all the data-driven approaches to classics out there nowadays…
Cheers, Chad
Very interesting @cb. Thank you. But I was wondering more about the latter: what percentage of nouns are 1sy, 2nd, 3rd etc declension… Not looking for a scientific answer, just an approximation, if someone would venture a guess.
I can’t tell you the percentage, but as you go up the declension ladder, the nouns become less frequent. I would say that by far the most nouns in Latin are first and second declension, more first than second, but a healthy helping of both. Quite a few third declension as well, but less than the first two. Of 4th and 5th declension nouns, considerably less, only really a handful of each, but most of them tend to be high frequency words.
@Barry Hofstetter
Thanks. It is helpful and comforting to know this: that I have kind of traversed the mid-way mark. I am progressing slowly but surely through Familia Romana, and have become very comfortable with Declension 1 and 2 nouns through the continual repetition in the text and exercises. I had taken a look ahead at Declension 3 noun paradigms and become a bit distressed. 
Verbs are the next big challenge.
Counting lemmas in LEMLAT 3.0 (only basis lexicon):
1: 22.6%
2: 33.6%
3: 38.9%
4: 4.4%
5: 0.5%
(22,218 nouns in total)
Similar frequencies are demonstrated in this topic: https://latin.stackexchange.com/questions/13642.
An important commentary from the above-mentioned thread: “5th declension is tiny in terms of lexis, but dies and res are pretty commonly used words”. So, these numbers should be normalized to the “frequency of occurrence”.
This is pretty interesting. 4th and 5th don’t really surprise. But like @lupusinfabula, I would have expected 1st and 2nd to top 3rd declension nouns. That surprised me 
This is just straight count of nouns in the lexicon? I wonder how it plays out in actual usage. I could imagine that 1st and 2nd declension nouns are more frequently used than 3rd declension for example.
Just a quick scan of a few passages of Livy suggests otherwise. There were about 3 times as many 3rd declension as 1st and 2nd (despite more repetition of 1st and 2nd declension words) in those passages.
The percentage of nouns in each declension has a lot to do with derivational endings. -tio inflates the third declension, and deverbatives in -us pump up the fourth. There’s just a handful of non-deverbative nouns in the fourth declension. Other endings inflate the first and second declensions. So I’m not sure why these statistics would be relevant or interesting.
I think you offer one of the reasons why this sort of statistic is interesting, Hylander - the declensions appear to admit different kinds of derivational forms, some of which are more common than others. That might say something about how Latin vocabulary grew over time, particularly if we were to take statistics for corpora at different stages of Latin.
It also appears to be of some paedagogical help - Lupus has taken encouragement that if he can master 1st, 2nd and 3rd declensions, then he will have crested the hill of noun morphology. I can certainly sypathise with that feeling!
Like theoldlove, I think the stats for actual usage would be a worthwhile next step, and they could be somewhat different.