Sound changes - complex or simple?

[Jonlang]

Would anyone here care to share the sound changes from their proto-lang to their target conlang? I ask because mine are currently taking up approximately 8 A4 pages and working words through them can be quite time consuming and I’m starting to think I may have over complicated it, but maybe not, maybe it’s supposed to be complicated

[Nortaneous]

It's supposed to be complicated. I don't think any of mine are complicated enough. Most overviews of sound changes I think aren't comprehensive enough - English is an exception, but English is fairly conservative in its consonants and the vowel shifts are typically made to seem a lot more random than they were.

[bradrn]

Most overviews of sound changes I think aren't comprehensive enough

Do you know of any better ones?

English … vowel shifts are typically made to seem a lot more random than they were.

How so?

[Nortaneous]

English … vowel shifts are typically made to seem a lot more random than they were.

How so?

To take one example, the TRAP-BATH and LOT-CLOTH splits were gap-filling by lengthening - there were /æ ɒ/ but no /æː ɒː/, and there was a general opposition of quantity elsewhere. (There was /a:/, which was at the time marginal and was later enlarged in a number of different ways in the various dialects, and probably /ɔː/, although it's possible that the lengthening of /ɒ/ was just driven by symmetry with the lengthening of /æː/.)

A good overview is one that thinks about the pressures and contradictory forces that give rise to systemic change. Thinking of it in terms of dialectics could work pretty well, although the generators or motive forces of the thesis and antithesis don't have to disappear after the synthesis. In English, there's the generator of symmetry, the generator of low vowel fronting, and the generator of quality-quantity correlation (i.e. there want to be more long than short vowels), and these can produce contradictions. EMidE æ a ǣ ā > a a ɛː ɔː produced a contradiction (no long low vowel) which was resolved by open syllable lengthening, and so on. Different languages can have different generators, but some generators are pretty common..

This doesn't get you to a full explanation of the GVS, but long high vowel breaking is common in West Germanic for whatever reason, and IIRC /i: u:/ > /ij uw/ or something in that ballpark has been used to explain the littera rule in Latin.

[bradrn]

English … vowel shifts are typically made to seem a lot more random than they were.

How so?

To take one example, the TRAP-BATH and LOT-CLOTH splits were gap-filling by lengthening - there were /æ ɒ/ but no /æː ɒː/, and there was a general opposition of quantity elsewhere. (There was /a:/, which was at the time marginal and was later enlarged in a number of different ways in the various dialects, and probably /ɔː/, although it's possible that the lengthening of /ɒ/ was just driven by symmetry with the lengthening of /æː/.)

A good overview is one that thinks about the pressures and contradictory forces that give rise to systemic change. Thinking of it in terms of dialectics could work pretty well, although the generators or motive forces of the thesis and antithesis don't have to disappear after the synthesis. In English, there's the generator of symmetry, the generator of low vowel fronting, and the generator of quality-quantity correlation (i.e. there want to be more long than short vowels), and these can produce contradictions. EMidE æ a ǣ ā > a a ɛː ɔː produced a contradiction (no long low vowel) which was resolved by open syllable lengthening, and so on. Different languages can have different generators, but some generators are pretty common..

This doesn't get you to a full explanation of the GVS, but long high vowel breaking is common in West Germanic for whatever reason, and IIRC /i: u:/ > /ij uw/ or something in that ballpark has been used to explain the littera rule in Latin.

This is actually pretty similar to how I think about sound change already. e.g. I have a conlang with an asymmetric vowel system /a aː ə i iː u/, with non-contrastive /uː/ in ideophones only. Presumably some descendents will just innovate /uː/, but the one I’m thinking about now resolves it by turning the quantity distinction into a quality distinction, /a aː i iː/ → /æ ɑ ɪ i/. But vowel’s don’t like to be too close together, so there is then some subsequent movement giving ultimately /a aː ə i iː u/ → /æ ɔ ə e i u/. And then I introduce a conspiracy to monosyllabic and sesquisyllabic roots, which includes posttonic vowel deletion, with umlaut onto the previous vowel (à la northern Vanuatuan) giving /æ a ɔ e ə o i u/. And so on and so forth.

Of course, it’s entirely possible to just list the sound changes one after another. But I find that rather unsatisfying. It’s much more interesting to look at the motivations of sound changes as well. (Taking into account of course that some changes do just occur with little motivation.)

[Ares Land]

Would anyone here care to share the sound changes from their proto-lang to their target conlang? I ask because mine are currently taking up approximately 8 A4 pages and working words through them can be quite time consuming and I’m starting to think I may have over complicated it, but maybe not, maybe it’s supposed to be complicated

Sure.

I don't claim to have any special deep insights, but I like working with diachronics. I find it's a bit of an art. It can be either simple or very elaborate. You can do very nice things with very simple changes.

Last year I worked on Voigari (basically, an alternate history Italian) and I really liked how it turned out.

Using it as an example feels incredibly conceited, but hey, people seemed to like the language, and I think it'll be easier to follow as it fits Romance diachronics.

A few samples:
Si non fia le Cesar, fia un'artista.

Si demadda padressu
Què resposa voi dellu?
Misa la manno esso petto
E cendo lebre i dedde

Code: Select all

[ei] → j /C_V
[ou] →w /C_V

ai→ɛ/ [+stressed]
e→ɛ/ [+stressed]
oi→ɛ/[+stressed]
o→ɔ/[+stressed]
e→ɛ/[+stressed]
ai→e/_
oi→e/_
i,ē→e/_
ī→i/_
au →o/_
[uō] →o/_
ū → u/_
[aā]→a/_


[mn]→/_# [-stress]
ns→n/_#
nt→n/_#
ns→s/_

bdg → vðɣ /V(")_V

--clusters
[kpgb]→t/_t
[kpgb]→t/_t
[kpgb]→s/_s
l→j/[stop]_


--palatals
skj→ʃ/_
stj→ʃ/_
ssj→ʃ/_
g→dʒ/_[jieɛ]/g_
k→tʃ/_[jieɛ]/k_
j→/[ʃʒ]_

t→ts/_[jieɛ]V
d→dʒ/_[jieɛ]V
kw→k/_

nj→ɲ/_
gn→ɲ/_
lj→ʎ/_
ll→l/_
sj→j/_/C_

This bit here follows SCA2 syntax. In plain English vowel are raised/broken whenever there's a final high vowel.
It's an exceedingly simple change, and it's pretty much the one bit that gives the language its character.
(Something similar happened in Old Tuscan, though it evolved differently; you'll find similar changes all over Central/Southern Italy)

Code: Select all

--Metaphony
ɛ→je/_CV…[iu]
ɔ→wo/_CV…[iu]
e→i/_…[iu]
o→u/…_[iu]

That bits is a feature of the Romanesco dialect.

Code: Select all

l→j/_[stop]


-- Finals i
t → ∅ /s_#

This part is what happened in Italian:

Code: Select all

s→i/V_# [+stressed]
as→e/_#
is→i/_#
es→i/_#

This is, on the face of it, a very simple change; except it gives the languages a very distinct esthetic and allows it to keep a tiny bit more nominal inflection.

Code: Select all

os→u/_#

That is standard Italian:

Code: Select all

--Final consonants
[tkdmns]→/_#

And that is specifically Southern:

Code: Select all

[-voiced] → [+voiced] /V_V
[-voiced] → [+voiced] /_r

Standard Italian:

Code: Select all

v→ ∅ / V_V
ðɣ → ∅ 

That is typical Southern Italian. Again, a very simple change that gives the language a particular feel.

Code: Select all

m→b/_b
n→d/_d
n→g/_g

These are pretty much the Latin to Italian sound changes, by the way, and all the changes there are attested somewhere in Central / Southern Italy. So you get an idea of what kind of changes to expect.

Index Diachronica is an unvaluable reference, but I agree with Nortaneous: you won't get the full story. For that you need a sizeable volume, focussing on one specific language.
Not necessarily English, but yeah, if you want anything exhaustive or definitive, you'll probably end up with the treatment of an European language. They're the best studied and their evolution is documented by centuries of written records. There must be some great stuff out there on Dravidian and Indo-Aryan languages, too.

On the time consuming part: I use the SCA² myself, and there are any number of sound change appliers out there. It's generally less time consuming, and anyway when I do the sound changes by hand I always end up doing something wrong. I believe taking some time figuring out how to make the SCA of your choice doing what you want would be time well spent.

[Raholeun]

Would anyone here care to share the sound changes from their proto-lang to their target conlang? I ask because mine are currently taking up approximately 8 A4 pages and working words through them can be quite time consuming and I’m starting to think I may have over complicated it, but maybe not, maybe it’s supposed to be complicated

To me it seems you are asking two questions:
- How many sound changes are realistic?
- How do you manage the complexity involved with a non-negligible amount of changes?

How many sound changes you want to have is mostly dependent on your preferences. If you are just looking for the illusion of naturalism in your conlang, the amount of sound changes does not need to be 8 pages in my opinion. However, some do like the challenge of simulating 10.000 years of history. As my conlang Sataw is supposed to be derived from a language very much like Proto-Austronesian, I knew more or less what level of deviation I should try to simulate by looking at other supposedly related languages. These sound changes fill up half a notebook, but after a thorough review and, more importantly, generalizing the changes into more general rules, they fit on a single page.

When it comes to managing those sound changes, I quickly found out some flexibility is needed. The sound change applier-strategy was simply too mechanical, and did not allow for any flexibility. Natural language development however is very fluid, and not at all as mechanical as some conlang descriptions make it seem. There are tons of exceptions to sound changes, and only in a limited number of priviliged languages can those exceptions be identified as coming from loan words. But in many understudied languages, we do not know the origin of the loan: the source language has no speakers left, there were borrowings from dialects or sociolects that are not present in the modern situation, etcetera. That is why I allow myself sometimes to just say that the change seems irregular and cannot be fully explained.

[Pabappa]

Would anyone here care to share the sound changes from their proto-lang to their target conlang? I ask because mine are currently taking up approximately 8 A4 pages and working words through them can be quite time consuming and I’m starting to think I may have over complicated it, but maybe not, maybe it’s supposed to be complicated

i'd like to see *your* list if youre willing to share. the way you describe it, i dont know if this is a lot or a little, because i dont know how densely youve typed, or how far apart in time the two endpoints are.

i would guess my languages tend to average about one sound change every 50 to 100 years, and i think that on average they are more conservative than the average rate of change on Earth, but not by much. the sound change lists for Poswa and Pabappa, my two best-developed languages, are 61 and 74 lines long, respectively.... this is for a timespan of 4,600 years. Despite these languages' distinctive sound, they change at about the same rate as the other languages on the planet, and so I think they're a good model to follow.

Right now, my list for Moonshine is 62 lines long and covers just under 3,000 years, but Moonshine is a much more rapidly changing language and in fact I've left a lot off the list since I've memorized enough of what goes on to do the changes in my mind.

I don't use sound change notation, .... I just prefer to write the changes out in plain language .... so I dont know if posting my lists will help all that much, but here's one of them, for a minor but important project called Play: http://www.frathwiki.com/Babakiam/Sound_changes . (I also use color highlighting, so it is much easier to read for me than for anyone else.)

[cedh]

I typically design my diachronic developments roughly as follows (this is not a fixed methodology, but more a summary of my experience of what tends to work):

• I usually tend to simulate a time interval of 1000-1500 years.
• I often set a "break point" somewhere in the middle, in order to have an intermediate stage of the language available (a) for a more realistic network of loanwords into and from other languages (these words will then undergo only the first or the second half of the changes before or after being borrowed), and (b) for creating dialects.
• This intermediate stage can also help with figuring out morphological changes, but I will usually not sketch out the intermediate grammar in more detail than just a few tables of inflectional paradigms and a handful of notes about syntax.
• In each of the two phases, I mix sound changes (or groups of related sound changes) with different purposes:
  • 2-3 "disruptive changes" that have a strong effect on the phoneme system (e.g. introducing new positional allophones and then removing the environment) and/or on the morphology (e.g. merging two segments that previously signaled a morphological distinction, or deleting word-final consonants)
  • 2-3 "flair changes" that affect the look and feel of the language, often without upsetting the grammar very much at all (e.g. unconditional chain shifts, or vowel diphthongizations/monophthongizations)
  • 1 "rhythm change" (e.g. stress shift, deletion of unstressed vowels, deletion of weak consonants...)
  • 5-10 "clean-up changes" to simplify some of the complexity brought about by the other sound changes, so that the language doesn't become unpronounceable at any point.

You can see some examples here:

• Ndak Ta to Ndok Aisô (elaborate dialect sound changes; note that the changes to Pre-Ndok and to the Southern dialect were designed by Dunomapuka several years earlier. Also cf. this description of the dialects for some example words.)
• Proto-Western to Gezoro (elaborate dialect sound changes; note that most of the Proto > Early changes were designed by Kodé several years earlier. Also cf. the lexicon (where "word" = Early Gezoro) for some examples of dialect words.)
• Proto-Coastal-Western to Doayâu
• Proto-Lukpanic to O Ayōndui
• Fáralo to Cəssın

[Jonlang]

i'd like to see *your* list if youre willing to share.

I would but I'm in the middle (more like the beginning) of revising much of it because what I want has changed since I first did it and I've tried to make some things more realistic than I had before, so I'm not really able to share anything at the moment. It was coming back to this process that made me think if I should ask about this.

However, I do appreciate everyone's help.

[Vardelm]

I typically design my diachronic developments roughly as follows (this is not a fixed methodology, but more a summary of my experience of what tends to work):

• stuff

Although I haven't looked at the examples, this is really helpful IMO. When I eventually get to developing descendants, I had though to go in roughly 500 year increments and base the number of sound changes on what would very roughly approximate categories of slow, medium, or rapid change. This post should act as a good guide for that.