There are other threads like this out there, but we don't have one here. I've also heard murmuring in #isharia about making a database of sound changes, which probably won't go anywhere in the future, but so long as we're talking about it we may as well have a thread.
Arapaho is probably the most unusual Algonquian language, diachronically. Here are some of the more notable changes from Proto-Algonquian to Arapaho:
Ojibwe voiced all obstruents except those in clusters (clusters then simplified).
Galician had g -> ħ (gheada).
Arapaho is probably the most unusual Algonquian language, diachronically. Here are some of the more notable changes from Proto-Algonquian to Arapaho:
*k is deleted, unconditionally, in all environments, leaving no trace.
*p becomes *k to close the resulting gap.
*č and *θ (probably /θ/, but also possibly /ɬ/) merge to *θ. Note that Arapaho keeps *θ and *l separate, which most Algonquian languages don't.
Ojibwe voiced all obstruents except those in clusters (clusters then simplified).
Galician had g -> ħ (gheada).
yeah weren't you the fuck who was convinced we needed grand comparative dictionaries, and now when you're doing something actually useful the best format you can think of is an anthologica forum thread?
Done. Each family gets its own spreadsheet. Currently, it's set up in seven columns:
The first column holds the proto-language. Originally, I just arranged it by extant language, but this makes it unclear whether (say) a Sanskrit change describes the fate of a sound in PIE or Proto-Indo-Iranian.
The second column holds the target language.
The third column holds the sound in the proto-language that underwent a change.
The fourth column holds the resulting sound.
The fifth column contains the environment.
The sixth column is used for citations.
You can use a seventh column for notes.
A short IRC discussion about some problems I've run into with regards to changes from Arapaho, and what this means for the spreadsheet.
————————————————
[10:11] <dhok> ok so
[10:11] <dhok> everyone listen up
[10:11] <dhok> (h13 you too)
[10:12] <dhok> I have just finished spreadsheeting the sound changes found in a paper about Arapaho and Atsina
[10:12] <dhok> it has been a wealth of information
[10:12] <dhok> and also indicates some problems we may run into
[10:18] <YaaliAnnar> Go on
[10:23] <dhok> well
[10:23] <dhok> first is the question of representation: do we use IPA for everything?
[10:23] <dhok> actual /y/ is nonexistant in the americas and most Americanists use <y> for /j/
[10:24] <dhok> secondly is the problem of exceptions
[10:24] <dhok> for example
[10:24] <dhok> in Arapaho
[10:24] <dhok> there was a change i -> u / oC_
[10:24] <dhok> except
[10:24] <dhok> if that C was /t θ n/ (ie a dental)
[10:24] <dhok> this didn't happen
[10:25] <dhok> thirdly we have the problem of making chain and related shifts clear
[10:25] <dhok> arapaho had Proto-Algonquian *k -> Ø and then *p -> k
[10:25] <dhok> clearly a chain shift
[10:25] <dhok> there were also a number of related shifts
[10:25] <Slereah> Do any hunter gatherers do their gathering in urban environments
[10:26] <dhok> this gets to the fourth and hariest problem
[10:26] <dhok> proto-arapaho-atsina had a vowel inventory of just three vowels: /i e o/
[10:26] <dhok> where *i was front and high, *e was front and non-high, and *o was back and non-high
[10:26] <dhok> then there were a set of parallel consonant shifts in Arapaho and Atsina
[10:27] <YaaliAnnar> Any reason why you shouldn’t use IPA?
[10:27] <dhok> where in Arapaho the front vowels *i *e would change consonants in the same way and *o would do it in another way
[10:27] <H13> dhok, forum
[10:27] <dhok> while in Atsina, *e and *o grouped together as non-high vowels and *i did its own thing
[10:27] <H13> anthologica already suffers from everything basically going on in irc
[10:27] <H13> but for things like this
[10:27] <dhok> h13 I'll just copy the chat here
[10:27] <H13> persistent
[10:27] <H13> storage
[10:28] <dhok> the issue is that there were a number of non-changes that arguably should be counted as changes
[10:28] <dhok> for example, in Arapaho, Goddard lists a change *k -> k / _o
[10:28] <dhok> which doesn't look like a change at all!
[10:28] <dhok> BUT
[10:29] <dhok> *k was part of a class of sounds *k *m *š that got affected before vowels in various ways
[10:30] <dhok> so *k -> k / _o was the mirror-"change" to *k -> č / _[i e]
[10:30] <Caroline> So then why is it even listed
[10:30] <dhok> just like *m -> w / _o was the mirror-change to *m -> b / _[i e]
[10:30] <Caroline> Oh, like that.
[10:30] <dhok> bingo
[10:30] <dhok> there wasn't a phonetic change
[10:31] <dhok> but it's part of a wider sound shift and in that sense it is a sound change, it's what happened to *k before *o
————————————————
[10:11] <dhok> ok so
[10:11] <dhok> everyone listen up
[10:11] <dhok> (h13 you too)
[10:12] <dhok> I have just finished spreadsheeting the sound changes found in a paper about Arapaho and Atsina
[10:12] <dhok> it has been a wealth of information
[10:12] <dhok> and also indicates some problems we may run into
[10:18] <YaaliAnnar> Go on
[10:23] <dhok> well
[10:23] <dhok> first is the question of representation: do we use IPA for everything?
[10:23] <dhok> actual /y/ is nonexistant in the americas and most Americanists use <y> for /j/
[10:24] <dhok> secondly is the problem of exceptions
[10:24] <dhok> for example
[10:24] <dhok> in Arapaho
[10:24] <dhok> there was a change i -> u / oC_
[10:24] <dhok> except
[10:24] <dhok> if that C was /t θ n/ (ie a dental)
[10:24] <dhok> this didn't happen
[10:25] <dhok> thirdly we have the problem of making chain and related shifts clear
[10:25] <dhok> arapaho had Proto-Algonquian *k -> Ø and then *p -> k
[10:25] <dhok> clearly a chain shift
[10:25] <dhok> there were also a number of related shifts
[10:25] <Slereah> Do any hunter gatherers do their gathering in urban environments
[10:26] <dhok> this gets to the fourth and hariest problem
[10:26] <dhok> proto-arapaho-atsina had a vowel inventory of just three vowels: /i e o/
[10:26] <dhok> where *i was front and high, *e was front and non-high, and *o was back and non-high
[10:26] <dhok> then there were a set of parallel consonant shifts in Arapaho and Atsina
[10:27] <YaaliAnnar> Any reason why you shouldn’t use IPA?
[10:27] <dhok> where in Arapaho the front vowels *i *e would change consonants in the same way and *o would do it in another way
[10:27] <H13> dhok, forum
[10:27] <dhok> while in Atsina, *e and *o grouped together as non-high vowels and *i did its own thing
[10:27] <H13> anthologica already suffers from everything basically going on in irc
[10:27] <H13> but for things like this
[10:27] <dhok> h13 I'll just copy the chat here
[10:27] <H13> persistent
[10:27] <H13> storage
[10:28] <dhok> the issue is that there were a number of non-changes that arguably should be counted as changes
[10:28] <dhok> for example, in Arapaho, Goddard lists a change *k -> k / _o
[10:28] <dhok> which doesn't look like a change at all!
[10:28] <dhok> BUT
[10:29] <dhok> *k was part of a class of sounds *k *m *š that got affected before vowels in various ways
[10:30] <dhok> so *k -> k / _o was the mirror-"change" to *k -> č / _[i e]
[10:30] <Caroline> So then why is it even listed
[10:30] <dhok> just like *m -> w / _o was the mirror-change to *m -> b / _[i e]
[10:30] <Caroline> Oh, like that.
[10:30] <dhok> bingo
[10:30] <dhok> there wasn't a phonetic change
[10:31] <dhok> but it's part of a wider sound shift and in that sense it is a sound change, it's what happened to *k before *o
you can use whatever notation, so long as you say — probably in another column — what the notation being used is. the polite term for 'he made it up' is 'idiotic'
consider the use of featural information, e.g. o + [+dental]_ for ur example
if they're common enough, put in another column called something like CHAINS and give each chain an index — langname+number so that group would be ARAPAHO0 or something
endnotes
I've included tags for changes. The spreadsheet has been updated with all the changes included in Goddard 1974 on Arapaho and Atsina, except for changes in accentuation, which were not very well understood at the time of writing (and to my knowledge still aren't really). Take a look!
I guess you could use indices to indicate feed and bleed orderings, if that's the intention? I do like this idea though.
I want to write code to analyze sound changes to produce graphs of what changes are ordered relative to one another...
I want to write code to analyze sound changes to produce graphs of what changes are ordered relative to one another...