Ugh…who would've thought the historical phonology of a linguistic
isolate would be so painful?!
Nota bene, the presentation of the sound changes to Basque is a lot like the Vulgar Latin → Old Provençal craziness 2: Electric Boogaloo.
I keep Aquitanian and Basque separate as opposed to nested because reasons (possibly not good ones, either). The sources I've read seem to leave the question of whether Aquitanian was the direct ancestor of Basque or not not totally resolved. Because in this sense I guess I'm a bit of a splitter I've listed them separately. I assumed the capital letters in the reconstructions provided by Wikipedia were the fortis series, following from the convention adopted by Egurtzegi. Quoted passages are also from Egurtzegi unless noted.
Proto-Vasconic to Aquitanian
From Egurtzegi, Ander (2013), "Phonetics and Phonology", in
Basque and Proto-Basque. <
https://www.academia.edu/3570162/2013a_ ... _Phonology>; Owstrowski, Manfred, "History of the Basque Language" <
http://hisp462.tamu.edu/Classes/603/Lec ... ueHist.pdf>; Wikipedia contributors (2014), "Aquitanian langauge".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =609638407>; Wikipedia contributors (2014), "Basque language".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =610796497>; and Wikipedia contributors (2014), "Vasconic languages".
Wikipedia, the Free Encyclopedia. <
http://en.wikipedia.org/w/index.php?tit ... =607530415>
Note that these changes are very likely highly incomplete and may be incorrect at times. I'm also assuming what is recorded as <x> in Aquitanian is actually /ʃ/.
ś → {s(ː),ʃ} / _#
ś → s
s → ʃ / i_#
ts → ʃs? sː?
S[+ fortis] → S[- voice]ː (specifically, the source lists t[+ fortis] k[+ fortis] → t(ː) k(ː), both of the tokens with optional length suffixes and *aTa → atta, so I'm extrapolating)
n[+ fortis] → n(ː) / V_V ?
n[- fortis] n[+ fortis] → {n,r}(?) n
N → [+ POA] / _S
r[+ fortis] → ɾ / _#
r[+ fortis] → r
Fortis *L is of uncertain outcome, seems to have been written as <l> or <ll>
g → k / #_ (sometimes?)
There seem to have been a few (variant?) forms which possibly show height assimilation in vowels
Proto-Vasconic to Basque
From Egurtzegi, Ander (2013), "Phonetics and Phonology", in
Basque and Proto-Basque. <
https://www.academia.edu/3570162/2013a_ ... _Phonology; Owstrowski, Manfred, "History of the Basque Language" <
http://hisp462.tamu.edu/Classes/603/Lec ... ueHist.pdf>; Wikipedia contributors (2014), "Proto-Basque language".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =605488703>; Wikipedia contributors (2014), "Aquitanian langauge".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =609638407>; Wikipedia contributors (2014), "Basque language".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =610796497>; Wikipedia contributors (2014), "Iberian language".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =601317949>; Wikipedia contributors (2014), "Basque dialects".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =595514648>; Wikipedia contributors (2014), "Biscayan dialect".
Wikipedia, the Free Encyclopedia. <
http://en.wikipedia.org/w/index.php?tit ... =613190357>; Wikipedia contributors (2014), "Gipuzkoan dialect".
Wikipedia, the Free Encyclopedia. <
https://en.wikipedia.org/w/index.php?ti ... =606871281>; Wikipedia contributors (2014), "Vasconic languages".
Wikipedia, the Free Encyclopedia. <
http://en.wikipedia.org/w/index.php?tit ... =607530415>; and Wikipedia contributors (2014), "Navarro-Lapurdian dialect".
Wikipedia, the Free Encyclopedia. <
http://en.wikipedia.org/w/index.php?tit ... =601150726>
Pre-Proto-Basque may have had some stuff involving reduplication that ended up dropping the first consonant
fortis → aspirated / "in a prominent position"
fortis → [-voice] / _V[-stressed]
lenis → devoiced / "in a prominent word-initial position"
lenis → voiced fricative (approximant? probably 12th-Century or later) / unstressed
lenis (voiced) → fricative / {l,r,s̻,s̺,V}_{l,r,s̻,s̺,V} (though sequences of *ld remained)
voiced fricative → [- voice] / C[+ sibilant]_
At least one reconstruction seems to indicate *s and *ś, which may have been an affricate and /s/. Pretty reliably, *-s tends to turn into -ts̺, and *-ś → *-ts̻, probably after the below-mentioned affrication. Beyond that, it's messy. *-tso seems to have become -tʃo/-tʃu, though.
Sʰ → F → h (→ Ø) / #_
S → S[+ voiced] / #_
"two similar vocalic segments" usually contract, except in some dialects (notably Biscayan)
Vn → Ṽ / _# (seems to have been reverted in most dialects, except for Souletin)
Vn → Ṽ / _V ?
V → Ṽ / _N (Souletin)
d → l / #_ (apparently not in verbs)
n → m / u_V
n → ɲ / {i,ɪ}_V
n → h̃ / V_V
N → [+ POA] / _C
nb → mː → m
b → m / _VN
{ɾ,r} → Ø / #_
l → ɾ / V_V
ɾ → r / _C
r → ɾ / _#
C → Ø / r_
Cr → Cɾ → CVɾ (maybe not perhaps a sound change per se, just a historical tendency)
ɾ → Ø / V_V (Souletin)
*L (fortis) → l (or lh → lː which then lost gemination?)
ð → ɾ (some sort of tap) (Biscayan, Guipuscoan, High Navarrese)
b → Ø / #_B (few exceptions, usually involving /u/)
F[+ voiced] → Ø / V_V (sometimes, usually involving "compound surnames"?)
S[+ voice] → [- voice] / F[+ sibilant]_
Ṽ → Vɲ / _V (not Souletin)
Ṽ → {Vn,diphthong} (not Souletin)
h̃ → h (not Souletin)
u ũ → y ỹ / _r(p(ʰ),β,k(ʰ),ɣ,l,s̺,s̻,ʃ,h) (Souletin)
u ũ → y ỹ / _{s̻,ts̺,ts̻} (but not _s̺) (Souletin)
õ → ũ (in Souletin)
Ø → a / #_{ra,ro} (sporadic)
Ø → e / #_r
Ø → e / #_{s̺,s̻}C
i → u / _(C…)u (Roncalese)
i → y / _(C…)y (Souletin)
e → o / _(C…)o (eastern dialects, Bermeo Biscayan)
e → o / o(C…)_ (eastern dialects)
a o e → ɛ u i / {i,u}(C…)_ (this [ɛ] is tentatively marked as such because the source transcribes it <e>, but says that it's not as close)
o → u / _n{C,#} (Souletin; some raising occurred elsewhere)
a → e / _$a (Biscayan, Alavese, some Guipuscoan)
o e → u i / _$a (raising of /o/ is less common)
e → i / _{n,C[+ sibilant]} (sporadic)
e → a / X_r ("mainly in the western dialects")
u i → o e / _r{C,#}
"some variations between /a/ and /e/ or /e/ and /i/" / _l{C,#}
Ø → j / V_{N,s̺,s̻}S
Ø → j / u_V (eastern dialects)
{w,y} → Ø / _ja
Ø → m / o$_V (Orozko Biscayan)
Ø → V / Vk_# (Zeberio Biscayan)
e → Ø / #_ (Navarrese, rare)
a → Ø / _V
V → Ø / Vj_
h → Ø (in western dialects; dialects such as Navarrese and Lapurdian retain it)
{w,β}h → f
*h may have metathesized given that it's usually only found in the first two syllables of proto-forms
h…h → Ø…h (says it "affect[s] both the oral /h/ and the nasalized aspiration")
*-ɾ → -h stuff in compounds
l n → ʎ ɲ / E_
{ɾ,r} → ʎ / {j,i}_ (eastern dialects)
s̻ ts̻ → ʃ tʃ / {E,j}_ (mostly Biscayan)
t → c / {E,j}_ ("some areas")
t → tʃ / {E,j}_ (partially spread amongst Biscayan and Guipuscoan)
d → ɟ (→ ʝ where voiced stops lenite) / {E,j}_? ("in some dialects")
d → ɟ (→ ʝ where voiced stops lenite) / {ʎ,ɲ}_ (Guipuscoan, High Navarrese)
{g,ɣ} → {ɟ,ʝ} / {E,j}_
g → dʒ / {E,j}_ ("in some Biscayan areas")
j → ʝ → j (northern High Navarrese, most Labourd, some Biscayan)
j → ʝ → ʒ (Souletin, sporadic in northwestern Biscayan)
j → ʝ (some Biscayan and Navarrese)
j → ʝ → ɟ (typical of Low Navarrese)
j → ʝ → ʒ → dʒ (northwestern Biscayan)
j → ʝ → ʒ → ʃ (Aescoan, Salazar, Roncalese, most of Southern High Navarrese)
j → ʝ → ʒ → ʃ → x (Guipuscoan, northwestern High Navarrese, eastern Biscayan)
j → χ (probably via intermediates, Wikipedia doesn't go into it very much)
ʎ ɲ → jl jn ("common in Low Navarrese, Labourdin, and is even regular in the High Navarrese of Sakana")
Syncope rules (Salazarese, Roncalese, High Navarrese, Low Navarrese, Aescoan):
— S_{ɾ,l} (more common in Roncalese and Salazar)
— C[+ sibilant]_ɾ (Roncalese and Salazar only)
— {ɾ,r}_O
— O_O
*N (fortis) → n
Something about final devoicing of stops and initial voiceless stops resulting from apocope
e → Ø / #Ur_
a → Ø / V_# (Guipuscoa; happens because of a reanalysis of the definite article)
V → Ø / _#, in trisyllables
i → Ø / _#, in disyllables
u {o,e} → Ø a / _#, in disyllables (only in the east)
"the -a or -e from the definite article" is dropped in Markina Biscayan and Getxo Biscayan
Some vowel metathesis which occurred only if the vowels matched each other in height
hu hi → ʊ ɪ / {o,e}_ (also happened with /a/ sometimes, but usually such sequences just dropped one vowel)
Something about diphthongs occurring where intervocalic /n/ was lost
VɪC → VCʲ
Glide dissimilation if the homorganic vowel was in the following syllable, but usually the glide just deleted
aɪ → eɪ → e (rare, but it sometimes happened)
aʊ → aɪ / ! _{ɾ,r,s̺,s̻} (Souletin, Roncalese)
eʊ → {e,egu}
eɪ → e / #_
oɪ → uɪ (rarely)
eð → j / #_V
e → j / #_a
e → Ø / #_e
ʊa → o "(especially after a velar stop)"
ʊe → e
C[- voice] → C[+ voice] / {l,N}_ (except Roncalese and Souletin)
Some speakers (Labourd and Low Navarrese?) have ʀ for r, and a few have ɢ̆ for ɾ
Something about historical final taps becoming aspiration in compounds
l → "velar" (some Souletin)
Souletin preserves something involving historical pretonic aspiration, apparently
Souletin (again!) keeps initial ʃ- and tʃ- distinct; Labourdin has ʃ-, the rest apparently have only tʃ-?
C[+ sibilant] → C[+ affricate] / _#
s̻ ts̺ → s̺ ts̻ (Biscayan, partially in Guipuscoan, Donostia, San Sebastián)
s̻ → s̺ / _{C,#} (sometimes)
From the Wikipedia article on Biscayan: "Convergence of sibilants: z, x and s > x, s and tz, tx and ts > tz." Good luck figuring that one out.
its̺ → tʃ / _# (Biscayan)
oa ea → u(e) i(e) / _#
Beterri Gipuscoan has VjV# where Biscayan has VɲV# and regular Basque just has VV#.
s̺ → tʃ / #_ (Gipuscoan)
ACCENT
- Navarrese and Labourdin seem to have gotten rid of phonemic accent; High Navarrese typically stresses the penult, while Low Navarrese and Labourdin are claimed to lack stress on the word level.
- Guipuscoa, southeastern Biscayan, and western varieties of Navarrese stress the second syllable (unless it is a disyllable, in which case the first syllable gets the accent, though a few varieties don't do this).
- North Biscayan does something with roots and affixes marked for prosody; "[m]ost native roots and almost all singular affixes are unaccented"; loans, "compounds and plural affixes" tend to be accented. Stress is typically assigned to the syllable immediately before the accent, but a few areas accent the penult or the antepenult.
- Souletin, unsurprisingly, does its own thing with accent. Stress usually falls on the penult, but contractions can mess with this (the example given in the paper is "alhába 'daughter' + abs. sg. -a > alhabá 'the daughter'"). Something similar is posited for "older. . .Salazarese". Roncalese was much the same, but the stress was stem-oriented as opposed to word-oriented unless contraction occurred, and there's some stuff about proparoxytones that Souletin didn't have.