Okay, so I'm having trouble figuring out the number of sound changes I want for my language. I've got ~45, but most of them are major, and almost all of them were created with the goal of shifting the language towards certain results (phonology, the evolution of a particular word, etc.) I want some more "minor" sound changes that don't really have a big impact on the language, but are nonetheless present. That said, I'm trying to figure out how many to add to each phase of my language's evolution. I've pretty much got the first phase of the language figured out, but the next two I'm not so sure on.
Basically, what I'm going for is:
The first phase, the one I've figured out, should have as many changes as there are between PIE and Proto-Italic. The number I'm going with for this phase is 29, based on the number of sound changes enumerated in Leiden and der Vaan's Etymological Dictionary of Latin; I have 18 so far.
The second phase should have as many changes as there are between Proto-Italic and Proto-Latino-Faliscan. I have 11 changes so far, and right now am planning 18 in total; however, this is hardly final, as it's based on the proportion of changes I have in the first phase and the number of changes I'm going for.
The third phase should have as many changes as there are between Proto-Latino-Faliscan and Classical Latin. I have 15 changes so far, and right now am planning on 24; though again, this isn't final since it's based on analogy between this phase and the first.
Is there any advice you can give on determining the number of sound changes for a language, and/or any good information on how many sound changes there are between Proto-Italic, Proto-Latino-Faliscan, and Classical Latin?
Honestly, I wouldn't worry all that much. Languages can change at wildly different rates, to the point where social factors such as massive social upheaval, contact and so on can be better predictors of change than simple time depth. Definitely throw in some more small-scale changes for extra naturalism but don't necessarily worry about sticking to a particular number.
I'd go further: sticking to a particular number is meaningless, because a) not all sound changes are equal, and b) sound changes are not countable.
On a), we can imagine two dialects of future English undergoing two changes each:
- in Alpha, /E:/ is shorted to /E/ before voiceless stops, and /I/ lengthens to /i:/ before nasals.
- in Beta, intervocalic voiced stops are dropped, and intervocalic voiceless stops become fricatives.
Alpha and Beta have both undergone 'two sound changes', but one has changed dramatically more than the other.
On b), how many changes were in the Great Vowel Shift? Do you count every starting vowel? Do you count the intermediate positions? Do you consider phonological realisations or phonemic systems? You could count the GVS as anything between 1 'change' (long vowels raise, but break if they are too high to raise further) and... what, 20-25 changes? Even just the outcome of /a:/ - is that one change (/a:/ > /e:/, with the diphthongal realisation of the latter in some dialects being a phonological detail) or is it four (/a:/ > /{:/ > /E:/ > /e:/ > /ej/)? Is the creation of the START and NORTH sets two changes (/arC/ > /a:rC/ and /QrC/ > /Q:rC/) or one (long vowels lengthen before coda /r/)?
What really happens is not "changes", but "change", a mass noun. Individual "changes" are largely a matter of notation. Different notations are used for different purposes. For instance, normally the raising/breaking of /a:/ is described as one sound change, from the point of view of the history of English. But if you're interested in the English of a given time period (for purposes of borrowings, for instance), then you're going to want to define it as a series of sound changes taking place sequentially over a century or two.
The best guide for what makes 'one' or 'two changes, in my opinion, is the presence of divergence - if two languages share one change but not another, that's a good reason to notate them as two changes, whereas if all descendents undergo both changes, it makes more sense usually to say they both undergo the 'whole of' the change rather than two different 'parts' of it in sequence.