Epenthesis and intrusive vowels in Lunigiana: diatopy, dichrony and phonologization

Discover the world at Leiden University

Epenthesis and intrusive vowels

in Lunigiana:

diatopy, diachrony and phonologization

Edoardo Cavirani Giornata di Fonologia -Pisa - 28/04/2014


Structure of the presentation

• Introduction • Hypotheses/Assumptions

• Geography

• Dialectology

• Diachronic Phonology

• Data • Phonetics

• Phonology

• Discussion and provisional conclusions • Diachrony

• Phonologization

2/60


Introduction

Diatopy≈Diachrony≈Phonologization

• Carrarese and Pontremolese represent two different stages of

the same diachronic change(s)

• The Pontremolese epenthetic vowel was what Carrarese intrusive

vowel is

• The Carrarese unstressed N is what Pontremolese unstressed N was

• The (Pontremolese) epenthetic vowel is the phonologized

version of the (Carrarese) intrusive vowel

• The intrusive vowel is phonetic but not phonological

• The epenthetic vowel is both phonetic and phonological

3/60


Introduction

Intrusive vs. Epenthetic vowel (Hall 2006)

Intrusive vowel

• either [ə], a copy of a nearby

vowel, or influenced by the

adjacent consonants place

• if copied, the intervening

consonant is a sonorant or a

guttural

• generally occurs in heterorganic

clusters

• likely to be optional; highly

variable duration; its presence

depends on speech rate

• no ‘repairing’ function (SSG)

Epenthetic vowel

• fixed quality or copied from a

nearby vowel

• if fixed, it doesn’t have to be [ə]

• if copied, no restrictions on the

intervening consonants

• its presence does not depend on

speech rate

• ‘repairing’ function (SSG)

4/60


Introduction

The grammar architecture (Bermúdez-Otero 2014)

• Different modules: Phonetics and Phonology • (at least) two different representations

• a change can apply just in one of them (without the change in one [phonetic]

representation implying its existence in the other [phonology])

• The mapping between the two different representations is learned,

language specific (and not universal as proposed by the substance-free

approach) and synchronically arbitrary

5/60


Introduction

Geography

Pontremoli

Carrara

Lunigiana

Italy

6/60


Introduction

Dialectology

• Lunigiana dialects belong to the

Northern-Italian varieties group

(Rohlfs 1966; Loporcaro 2009)

• Peripheral varieties (Bertoni

& Bartoli 1925; Andersen 1988):

“likely to see the development of

elaborate phonetic norms and

the proliferation of low-level

pronunciation rules”

(Trudgill 1992: 206)

7/60


Introduction

Dialectology

• Three major groups

(Giannarelli 1913):

• Tusco-Ligurian

• Tusco-Emilian (Carr.)

• Liguro-Emilian (Pontr.)

• The linguistic borders

are not that sharp:

the Ligurian features decrease

from N, NE to S and the

Emilian ones from N to S,

where they both melt with

the Tuscan features

8/60


Introduction

Diachronic Phonology

unstressed vowel reduction/deletion

(shared with other NI dialects)

• syncope

SILVĀTICA(M) ‘wild f.’ > Carr. [səlˈvatka]

Pontr. [sarˈvadga]

• apocope

SILVĀTICU(M) ‘wild m.’ > Carr. [səlˈvatk(ə)]

Pontr. [sarˈvadg(ə)]

9/60


Introduction

Diachronic Phonology

unstressed vowel reduction/deletion

(micro-)diacronic evolution stages ≈ (micro-)diatopic variation

“The […] unstressed vowel deletion, which is a permanent feature

in the Emilian dialect, is (almost) normal in the Higher Magra

Valley (Pontremoli) […], while the more you descend the valley, the

more the deletion alternates with the reduction to the faint vowel

[schwa, E.C.], which is quite regular in a lot of Lower Lunigiana

dialects. As a result […] some scholars […] have been induced to

consider the presence of the faint vowel as the ‘joining link’

between the unstressed vowel persistence of the Tuscan dialect and

the constant deletion of the Emilian one” (Luciani 1999: 82)

10/60


Data

Phonetics and Phonology

• Two prosodic templates:

• paroxitones ˈσ.CCv]

• proparoxitones ˈσ.Cv1Cv2]

• Three (consonant cluster) sonority conditions:

• falling sonority contour (]w vocoid)

• even sonority contour (]w vocoid)

• rising sonority contour (I/E vocoids)

• Three main phrasal contexts:

• before a consonant-initial word ]w C

• before a pause ]w #

• before a vowel-initial word or a vocalic ending (]w) V

11/60


Data

Carrarese - Paroxitones ( SC)

F1 F2 v ms V ms v/V Transcr. F1 F2 v ms V ms v/V Transcr.

colpo 461 1632 69 61 113% [ˈkolpə] 0 0 0 192 0 [ˈkolp]

forno 369 1664 56 64 88% [ˈfornə] 0 0 0 189 0 [ˈforn]

merlo 507 1589 57 148 38% [ˈmɛrlə] 0 0 0 169 0 [ˈmɛrl]





forno 494 1336 74 43 172% [ˈfornə] 546 1427 94 156 60% [ˈfornə]

merlo 0 0 0 101 0 [ˈmɛrl] 0 0 0 180 0 [ˈmɛrl]



merlo 546 1391 39 123 32% [ˈmɛrlə] 746 1642 84 188 45% [ˈmɛrl

ə]

colpo N N N N N N 0 0 0 113 0 [ˈkolp]

forno N N N N N N 0 0 0 162 0 [ˈforn]

merlo 439 1357 73 68 107% [ˈmɛrlə] 440 1435 120 207 58% [ˈmɛrlə]

Formsˈσ.CCv] C ˈσ.CCv] #

AC

BD

AM

ED

MV

12/60


Carrarese F1 F2 v ms V ms v/V

mean 469 1534 54 89 74%

SD 56 194 13 41 42%

mean 577 1501 99 184 54%

SD 156 122 19 26 8%

mean 491 1527 63 108 70%

SD 89 179 23 54 38%

ˈσ.CCv] C

ˈσ.CCv] #

Total

Data


The word-final vocoid of Carr. paroxitones

• is a schwa (F1 and F2 close to the schwa ‘ideal’ ones; Silverman 2011)

• shows a great F2 variation, especially in ]w C (coarticulation)

• displays pretty much the same values in both the phrasal contexts*

*if the form is followed by a vowel, the word-final vocoid is not realized and the

(phonologically) C]w is re-syllabified

word-final vocoid [ə]

/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

13/60


Data


The word-final vocoid of Carr. paroxitones

• is the 70% of the stressed vowel duration

• is proportionally shorter in the phrase-final context (54% vs. 74%) even if it displays a greater absolute duration (99 ms vs. 54 ms) due to the phrase-

final lengthening


mean 467 1506 54 89 74%

SD 52 195 13 41 42%

mean 577 1501 99 184 54%

SD 156 122 19 26 8%

mean 491 1527 63 108 70%

SD 89 179 23 54 38%

ˈσ.CCv] C

ˈσ.CCv] #

Total


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

14/60


Extremely low frequency

of v]w # realization:

3 out of 15 tokens (20%);

regular absence in (]w)V

Extreme shortness of some

v]w C realizations:

less than the 40% of the

stressed vowel duration or

even unrealized

ə-like F1 and F2

ˈσ.CCv] vocoid = ə-like rel.

Data










forno 494 1336 74 43 172% [ˈfornə] 546 1427 94 156 60% [ˈfornə]

merlo 0 0 0 101 0 [ˈmɛrl] 0 0 0 180 0 [ˈmɛrl]



merlo 546 1391 39 123 32% [ˈmɛrlə] 746 1642 84 188 45% [ˈmɛrl

ə]

colpo N N N N N N 0 0 0 113 0 [ˈkolp]

forno N N N N N N 0 0 0 162 0 [ˈforn]

merlo 439 1357 73 68 107% [ˈmɛrlə] 440 1435 120 207 58% [ˈmɛrlə]


AC

BD

AM

ED

MV

15/60


Data

Pontremolese - Paroxitones ( SC)


colpo 393 1244 49 87 56% [ˈkurpə] 0 0 0 167 0 [ˈkurp]

forno 511 1292 135 111 121% [ˈfurnə] 0 0 0 179 0 [ˈfurn]

merlo 486 1179 94 165 57% [ˈmørlə] 0 0 0 211 0 [ˈmørl]

colpo 371 1506 41 37 111% [ˈkurpə] 369 1450 93 126 74% [ˈkurpə]

forno 419 1362 70 46 152% [ˈfurnə] 350 1493 100 143 70% [ˈfurnə]

merlo 446 1520 58 93 62% [ˈmørlə] 411 1451 37 80 46% [ˈmørlə]










MM

GB

AS

LB


DP

16/60


Pontremol. F1 F2 v ms V ms v/V

mean 432 1448 65 86 87%

SD 52 184 25 46 35%

mean 443 1492 90 184 74%

SD 69 65 29 19 23%

mean 437 1465 75 100 82%

SD 58 148 29 41 31%

ˈσ.CCv] C

ˈσ.CCv] #

Total

Data


The word-final vocoid of Pontr. paroxitones

• is a schwa (F1 and F2 similar to the ‘ideal’ and the Carr. [ə] ones)

• shows a great F2 variation in ]w C (coarticulation)

• displays pretty much the same values in both the phrasal contexts


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

17/60



mean 432 1448 65 86 87%

SD 52 184 25 46 35%

mean 443 1492 90 184 74%

SD 69 65 29 19 23%

mean 437 1465 75 100 82%

SD 58 148 29 41 31%

ˈσ.CCv] C

ˈσ.CCv] #

Total

Data


The word-final vocoid of Pontr. paroxitones

• is the 82% of the stressed vowel duration (vs. Carr. 70%)

• is proportionally shorter in the phrase-final context (74% vs. 87%) even if it displays a greater absolute duration (90 ms vs. 65 ms) due to the phrase-

final lengthening


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

18/60


Data




forno 511 1292 135 111 121% [ˈfurnə] 0 0 0 179 0 [ˈfurn]














MM

GB

AS

LB


DP

Constantly realized in ]w C;

optionally (LB and AS) or

regularly (DP) unrealized in

]w #; absent in (]w)V

Short duration of the ]w C

vocoids; extreme duration

variability

ə-like F1 and F2

ˈσ.CCv] vocoid = ə-like rel.

19/60


D

ata

Car

rare

se -

Pro

par

ox

ito

nes

v1 v2 v1 v2 v1 v2 V v1 v2 v1 v2 v1 v2 v1 v2 V v1 v2

tiepido 350 453 1827 1627 64 120 153 42% 79% [ˈtepədə] 464 491 1683 1584 83 85 186 45% 46% [ˈtep

əd

ə]

selvatico 662 404 1691 1490 51 48 76 67% 64% [səlˈvatəkə] 329 477 1152 1596 46 54 148 31% 36% [səlˈvatək

ə]

stomaco 461 424 1127 1380 43 39 57 76% 69% [ˈstoməkə] 0 0 0 0 0 0 Nc 0 0 [ˈstomk]

manico 599 440 1035 1667 43 50 103 42% 48% [ˈmanəkə] 0 0 0 0 0 0 Nc 0 0 [ˈmank]

tenero 456 0 1807 0 59 0 92 64% 0 [ˈtenər] 445 0 1663 0 86 0 157 55% 0 [ˈtenər]

giovane 369 0 1477 0 64 0 61 105% 0 [ˈdzovən] 429 359 1552 1653 43 68 193 22% 36% [ˈdzovən

ə]

libero 446 0 1545 0 41 0 122 34% 0 [ˈlibər] 490 0 1649 0 32 0 110 29% 0 [ˈlibər]

tiepido 0 0 0 0 0 0 Nc 0 0 [ˈtepd] 0 0 0 0 0 0 Nc 0 0 [ˈtepd]

selvatico 665 449 1438 1554 41 39 88 46% 44% [səlˈvatəkə] 612 0 1691 0 75 0 143 52% 0 [səlˈvatək]

stomaco N N N N N N N N N N 0 0 0 0 0 0 Nc 0 0 [ˈstomk]

manico 616 565 1948 1671 32 51 81 39% 64% [ˈmanəkə] 490 0 1475 0 32 0 152 21% 0 [ˈman

ək]

tenero 494 0 1992 0 47 0 141 33% 0 [ˈtenər] 501 0 2196 0 43 0 144 30% 0 [ˈten

ər]

giovane 490 0 1286 0 59 0 122 48% 0 [ˈdzovən] 0 0 0 0 0 0 88 0 0 [ˈdzovn]

libero 0 0 0 0 0 0 N 0 0 [ˈlibr] 0 0 0 0 0 0 Nc 0 0 [ˈlibr]

tiepido 413 433 1461 1557 50 67 189 27% 36% [ˈtepəd

ə] 539 513 1482 1498 39 98 142 27% 68% [ˈtep

ədə]

selvatico 349 563 1919 1375 37 44 93 40% 47% [səlˈvatək

ə] 0 543 0 1556 0 61 149 0 41% [səlˈvatk

ə]

stomaco 492 495 1180 1223 34 62 60 56% 104% [ˈstoməkə] 504 486 1208 1553 69 43 97 71% 45% [ˈstoməkə]

manico 406 539 1351 1633 31 42 76 41% 55% [ˈmanəkə] 343 0 1675 0 63 0 130 49% 0 [ˈman

ək]

tenero 506 514 1518 1433 49 81 64 77% 128% [ˈtenərə] N N N N N N N N N N

giovane 476 0 1266 0 63 0 57 110% 0 [ˈdzovən] 548 466 1396 1396 32 50 173 18% 29% [ˈdzovən

ə]

libero 452 0 1455 0 63 0 50 127% 0 [ˈlibər] 440 0 1478 0 44 0 124 35% 0 [ˈlibər]

tiepido 452 432 1063 1247 24 22 101 24% 22% [ˈtepəd

ə] 0 0 0 0 0 0 Nc 0 0 [ˈtepd]

selvatico 320 444 1605 1135 33 39 68 48% 58% [səlˈvatəkə] 442 0 1989 0 50 0 106 47% 0 [səlˈvat

ək]

stomaco 620 0 1140 0 27 0 31 88% 0 [ˈstomək] 476 0 1134 0 44 0 97 45% 0 [ˈstomək]

manico 457 464 1451 1315 37 36 66 56% 54% [ˈmanəkə] 600 494 1476 1450 62 59 155 40% 56% [ˈmanəkə]


giovane 0 0 0 0 0 0 Nc 0 0 [ˈdzovn] 514 441 1348 821 57 71 130 44% 55% [ˈdzovənə]

libero 0 0 0 0 0 0 Nc 0 0 [ˈlibr] 0 0 0 0 0 0 Nc 0 0 [ˈlibr]

tiepido 375 436 1310 1322 70 67 81 86% 83% [ˈtepədə] 433 407 1397 1551 65 87 149 44% 58% [ˈtepədə]

selvatico 327 393 1450 1411 58 43 100 58% 44% [səlˈvatəkə] 320 401 1433 1390 81 108 156 52% 69% [səlˈvatəkə]

stomaco 0 415 0 1077 0 21 56 0 37% [ˈstomkə] 0 0 0 0 0 0 Nc 0 0 [ˈstomk]

manico N N N N N N N N N N 320 0 2239 0 57 0 173 33% 0 [ˈmanək]


giovane 385 0 1191 0 63 0 95 66% 0 [ˈdzovən] 409 440 1334 1438 69 106 181 38% 59% [ˈdzovən]

libero 453 407 1327 1338 50 69 166 30% 42% [ˈlibər

ə] 421 426 1246 1287 68 89 105 65% 83% [ˈlibərə]

AM

AC

BD

ˈσ.Cv1Cv2] #

F1 F2 Duration v/VTranscr.

ˈσ.Cv1Cv2] C

FormsTranscr.

F1 F2 Duration v/V

ED

MV

20/60


v1 v2 v1 v2 v1 v2 V v1 v2

mean 468 459 1458 1414 49 52 91 61% 60%

SD 97 54 265 180 15 23 38 31% 26%

mean 460 457 1559 1444 61 75 145 42% 52%

SD 81 51 299 213 25 22 31 15% 16%

mean 464 459 1504 1427 54 62 117 52% 57%

SD 90 52 283 192 21 25 44 27% 22%

v/V

ˈσ.Cv1Cv2] C

ˈσ.Cv1Cv2] #

Total

CarrareseDurationF1 F2

Data

Carrarese - Proparoxitones

The word-final vocoid of Carr. proparoxitones

• is a schwa (F1 and F2 similar to the ‘ideal’ and the paroxitone ones)


• shows a great F2 variation (coarticulation)

21/60


v1 v2 v1 v2 v1 v2 V v1 v2

mean 468 459 1458 1414 49 52 91 61% 60%

SD 97 54 265 180 15 23 38 31% 26%

mean 460 457 1559 1444 61 75 145 42% 52%

SD 81 51 299 213 25 22 31 15% 16%

mean 464 459 1504 1427 54 62 117 52% 57%

SD 90 52 283 192 21 25 44 27% 22%

v/V

ˈσ.Cv1Cv2] C

ˈσ.Cv1Cv2] #

Total

CarrareseDurationF1 F2

Data


The word-medial vocoid of Carr. proparoxitones


• is pretty much the same in both the phrasal context

• shows a greater (F2) variation with respect to v2 (coarticulation)

22/60


Data


Word-medial and -final vocoids of Carr. proparoxitones

• both display schwa-like values

• greater variation of v1 with respect to v2 (coarticulation)

word-medial vocoid [ə]


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

23/60


Data


The Carr. proparoxitone vocoids

• are about half of the stressed vowel duration

• are proportionally shorter in ]w # even if they display the phrase-

final lengthening effect (greater absolute durations)

V

ms v/V ms v/V ms

mean 49 61% 52 60% 91

SD 15 31% 23 26% 38

mean 61 42% 75 52% 145

SD 25 15% 22 16% 31

mean 54 52% 62 57% 117

SD 21 27% 25 22% 44

v1 v2

v]w C

v]w #

Total

24/60


• v2 is often unrealized: 61% in ]w # and 46% in ]w C

• v1 can be unrealized as well: 30% in ]w # and 16% in ]w C

• they are about the half of the stressed vowel duration

• they display schwa-like values

• they show a great F2 variation

ˈσ.Cv1Cv2] vocoids = ə-like releases

Data


25/60


D

ata

Po

ntr

emo

lese

-

Pro

par

ox

ito

nes

(

SC

)


tiepido 297 389 877 1024 94 101 155 60% 65% [ˈteʋdə] 345 379 920 1514 125 62 151 83% 41% [ˈteudə]

selvatico 0 391 0 1207 0 74 98 0 76% [sarˈvadgə] 0 386 0 1878 0 53 176 0 30% [sarˈvadgə]

stomaco 0 411 0 1227 0 123 81 0 152% [ˈstumgə] 0 415 0 1566 0 129 109 0 119% [ˈstumgə]

manico 0 466 0 1469 0 93 105 0 88% [ˈmankə] 0 416 0 1748 0 79 122 0 65% [ˈmankə]

tiepido 0 412 0 1593 0 88 81 0 108% [ˈtevdə] 0 413 0 1584 0 85 85 0 100% [ˈtevdə]



manico 0 380 0 1596 0 61 53 0 116% [ˈmangə] 0 389 0 1552 0 75 149 0 50% [ˈmangə]













ˈσ.Cv1Cv2] C

Forms Duration v/VTranscr.

ˈσ.Cv1Cv2] #

MM

F1 F2

GB

DP

LB

AS

Duration v/VTranscr.

F1 F2

26/60


v1 v2 v1 v2 v1 v2 V v1 v2

mean 0 406 0 1496 0 72 86 0 86%

SD 0 43 0 213 0 24 29 0 24%

mean 0 413 0 1576 0 88 121 0 76%

SD 0 41 0 132 0 32 28 0 31%

mean 0 409 0 1536 0 80 103 0 81%

SD 0 42 0 179 0 29 33 0 28%

Pontremol.F1 F2 Duration

ˈσ.Cv1Cv2] C

ˈσ.Cv1Cv2] #

Total

v/V

Data

Pontremolese - Proparoxitones ( SC)

The word-final vocoid of Pontr. proparoxitones



• shows a great F2 variation, especially in ]w C (coarticulation)

27/60


Data


Word-final vocoids of Pontr. proparoxitones

• display schwa-like values in both ]w # and ]w C

• greater variation of F2 with respect to F1 (coarticulation)


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

28/60


v1 v2 v1 v2 v1 v2 V v1 v2

mean 0 406 0 1496 0 72 86 0 86%

SD 0 43 0 213 0 24 29 0 24%

mean 0 413 0 1576 0 88 121 0 76%

SD 0 41 0 132 0 32 28 0 31%

mean 0 409 0 1536 0 80 103 0 81%

SD 0 42 0 179 0 29 33 0 28%

Pontremol.F1 F2 Duration

ˈσ.Cv1Cv2] C

ˈσ.Cv1Cv2] #

Total

v/V

Data


The word-final vocoid of Pontr. proparoxitones

• is slightly longer in ]w C (86%) than in ]w # (76%) even if they display

the phrase-final lengthening effect (greater absolute durations)

• is pretty perfectly as long as the paroxitones one (81% vs. 80%)

29/60


• v2 is always realized (v1 is never realized)

• v2 is longer in ]w C (86%) than in ]w # (76%)

• they are shorter than the stressed vowel duration

• display schwa-like values

• show a great F2 variation

Prop. v]w vocoid = ə-like release

Data

Pontremolese - Proparoxitones

30/60


Data

Carrarese - Paroxitones ( SC > I)


libro 442 1470 58 49 118% [ˈlibər] 0 0 0 138 0 [ˈlibr]

magro 467 1767 50 97 52% [ˈmagər] 0 0 0 189 0 [ˈmagr]

quattro 580 1744 49 99 50% [ˈkwatər] 0 0 0 165 0 [ˈkwatr]


magro 0 0 0 120 0 [ˈmagr] 0 0 0 168 0 [ˈmagr]

quattro 0 0 0 190 0 [ˈkwatr] 0 0 0 94 0 [ˈkwatr]


magro 436 1421 53 96 55% [ˈmagər] 488 1420 59 218 27% [ˈmagər]

quattro 0 0 0 88 0 [ˈkwatr] 423 1463 60 134 45% [ˈkwatər]




libro 425 1207 143 76 189% [ˈlibər] 426 1324 75 124 60% [ˈlibər]

magro 426 1516 54 112 48% [ˈmagər] 427 1481 58 225 26% [ˈmag

ər]

quattro 444 1402 39 128 31% [ˈkwatər] 426 1359 46 118 39% [ˈkwat

ər]

Formsˈσ.CIC] C ˈσ.CIC] #

AC

BD

ED

AM

MV

31/60


Data


When the apocope creates a ]w consonant cluster with a rising

sonority contour an intrusive vocoid can be produced which:

• shows schwa-like F1 and F2 values

• doesn’t display a significant difference between ]w C and ]w #

• shows a greater F2 variation


mean 472 1458 65 99 79%

SD 51 183 31 33 48%

mean 438 1409 60 157 39%

SD 28 67 10 39 14%

mean 461 1442 63 128 66%

SD 46 153 25 46 43%

Total

ˈσ.CIC] #

ˈσ.CIC] C

/u/

I vocoid [ə]

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

32/60


Data



sonority contour an intrusive vocoid can be produced which:

• doesn’t undergo the phrase-final lengthening: only the stressed vowels do

• is (proportionally) extremely short (39%) in ]w #


mean 472 1458 65 99 79%

SD 51 183 31 33 48%

mean 438 1409 60 157 39%

SD 28 67 10 39 14%

mean 461 1442 63 128 66%

SD 46 153 25 46 43%

Total

ˈσ.CIC] #

ˈσ.CIC] C

/u/

I vocoid [ə]

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

33/60


Frequently unrealized (67%)

in ]w # and (less frequently:

33%) in ]w C

Short duration (usually

shorter than 60 ms)

ə-like F1 and F2

ˈσ.CvC] I vocoid = ə-like rel.

Data




magro 467 1767 50 97 52% [ˈmagər] 0 0 0 189 0 [ˈmagr]

quattro 580 1744 49 99 50% [ˈkwatər] 0 0 0 165 0 [ˈkwatr]





magro 436 1421 53 96 55% [ˈmagər] 488 1420 59 218 27% [ˈmagər]

quattro 0 0 0 88 0 [ˈkwatr] 423 1463 60 134 45% [ˈkwatər]




libro 425 1207 143 76 189% [ˈlibər] 426 1324 75 124 60% [ˈlibər]

magro 426 1516 54 112 48% [ˈmagər] 427 1481 58 225 26% [ˈmag

ər]

quattro 444 1402 39 128 31% [ˈkwatər] 426 1359 46 118 39% [ˈkwat

ər]

Formsˈσ.CIC] C ˈσ.CIC] #

AC

BD

ED

AM

MV

34/60


Data

Pontremolese - Paroxitones ( SC > E)


libro 564 880 91 160 57% [ˈlibɐr] 571 1033 105 137 77% [ˈlibɐr]

magro 616 1253 95 123 77% [ˈmagɐr] 628 1429 153 170 90% [ˈmagɐr]

quattro 541 1375 45 102 44% [ˈkwatɐr] 581 1390 83 154 54% [ˈkwatɐr]













GB

Formsˈσ.CEC] C ˈσ.CEC] #

DP

MM

LB

AS

35/60


Data



sonority contour an epenthetic vocoid is produced which:

• shows a slightly higher F1 and lower F2 ([ɐ])

• displays a greater F2 variation

• doesn’t show significant differences between ]w C and ]w #


mean 549 1386 67 90 77%

SD 51 187 18 26 21%

mean 564 1415 92 184 68%

SD 53 133 31 22 22%

mean 556 1400 79 113 72%

SD 52 160 28 33 21%

ˈσ.CEC] #

Total

ˈσ.CEC] C

/u/

/i/

E vocoid [ɐ]

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

36/60


Data



sonority contour an epenthetic vocoid is produced which:

• doesn’t display significant differences if occurring in ]w C or ]w #

• is longer (68%) with respect to Carr. I in ]w # (39%)

• undergoes the phrase-final lengthening


mean 549 1386 67 90 77%

SD 51 187 18 26 21%

mean 564 1415 92 184 68%

SD 53 133 31 22 22%

mean 556 1400 79 113 72%

SD 52 160 28 33 21%

ˈσ.CEC] #

Total

ˈσ.CEC] C

/u/

/i/

E vocoid [ɐ]

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

37/60


It is regularly realized in

both the phrasal contexts

(unlike Carr. I vocoid)

It is shorter than a stressed

vowel (but longer than Carr.

I vocoid)

It is lower and more back

than a schwa

ˈσ.CvC] E vocoid = /a/

Data


















GB

Formsˈσ.CEC] C ˈσ.CEC] #

DP

MM

LB

AS

38/60


D

ata

Po

ntr

emo

lese

-

Pro

par

ox

ito

nes

(

SC

> E

)


tenero 567 0 1219 0 56 0 87 64% 0 [ˈtenɐr] 660 0 1286 0 69 0 140 49% 0 [ˈtenɐr]

libero 547 0 1062 0 109 0 148 74% 0 [ˈlibɐr] 624 0 1225 0 107 0 121 88% 0 [ˈlibɐr]

giovane 423 0 690 0 140 0 171 82% 0 [ˈzuʊŋ] 420 0 692 0 166 0 145 115% 0 [ˈzuʊŋ]

asino 303 0 629 0 105 0 139 76% 0 [ˈazʊŋ] 345 0 745 0 81 0 178 45% 0 [ˈazʊŋ]

tenero N N N N N N N N N N 530 0 1487 0 70 0 145 48% 0 [ˈtenɐr]





libero N N N N N N N N N N N N N N N N N N N N

giovane N N N N N N N N N N 283 365 702 1368 102 95 81 126% 117% [ˈzuʊŋə]






tenero 532 438 1571 1608 44 53 81 54% 65% [ˈtenɐrə] 569 0 1610 0 91 0 105 86% 0 [ˈtenɐr]





F1 F2 Duration v/VF2

GB

Transcr.

DP

MM

LB

AS

FormsF1

ˈσ.CEC] C ˈσ.CEC] #

39/60


Data

Pontremolese - Proparoxitones ( SC >E)

The E vocoids of Pontr. proparoxitones

• either an [a]-like (aE, before a liquida) or an [u]-like (uE, before a

velar nasal) vocoid

• they show a greater F2 variation

• aE values of paroxitones and proparoxitones are extremely similar

aE v2 aE v2 aE v2 V aE v2 uE v2 uE v2 uE v2 V uE v2

mean 589 0 1371 0 71 0 96 73% 0 348 0 762 0 88 0 126 70% 0

SD 54 0 183 0 26 0 26 18% 0 61 0 130 0 30 0 33 17% 0

mean 605 0 1414 0 93 0 115 84% 0 320 0 784 0 109 0 146 81% 0

SD 77 0 139 0 23 0 22 26% 0 57 0 150 0 45 0 40 39% 0

mean 597 0 1394 0 83 0 106 79% 0 333 0 774 0 99 0 137 75% 0

SD 66 0 157 0 26 0 25 23% 0 59 0 137 0 39 0 37 30% 0

ˈσ.CEC] #

Total

aE

Pontremol.

uE

ˈσ.CEC] C ˈσ.CEC] C

ˈσ.CEC] #

Total

F2 Duration v/VF1 F2 Duration v/V F1

40/60


Data



• either an [a]-like (aE) or an [u]-like (uE) vocoid

• aE is significantly different from the ]w ə-like vocoid

• aE ellipsis overlaps with the unstressed /a/ one

• uE ellipsis overlaps with the /u/ one

uE

/u/

/i/

aE

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1


unstressed /a/

/u/

/i/

aE

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

41/60


Data



• aE and uE displays similar duration values in both the ph. contexts

• undergo the phrase-final lengthening

aE v2 aE v2 aE v2 V aE v2 uE v2 uE v2 uE v2 V uE v2

mean 589 0 1371 0 71 0 96 73% 0 348 0 762 0 88 0 126 70% 0

SD 54 0 183 0 26 0 26 18% 0 61 0 130 0 30 0 33 17% 0

mean 605 0 1414 0 93 0 115 84% 0 320 0 784 0 109 0 146 81% 0

SD 77 0 139 0 23 0 22 26% 0 57 0 150 0 45 0 40 39% 0

mean 597 0 1394 0 83 0 106 79% 0 333 0 774 0 99 0 137 75% 0

SD 66 0 157 0 26 0 25 23% 0 59 0 137 0 39 0 37 30% 0

ˈσ.CEC] #

Total

aE

Pontremol.

uE

ˈσ.CEC] C ˈσ.CEC] C

ˈσ.CEC] #

Total

F2 Duration v/VF1 F2 Duration v/V F1

42/60


They are regularly realized

in both the phrasal contexts

(differently from Carr.)

They are shorter than a

stressed vowel (but longer

than Carr. I vocoid)

Their F1 and F2 point to [ɐ]

and [u]

Pontr. E vocoids = /a/ & /u/

Data







tenero N N N N N N N N N N 530 0 1487 0 70 0 145 48% 0 [ˈtenɐr]





libero N N N N N N N N N N N N N N N N N N N N

giovane N N N N N N N N N N 283 365 702 1368 102 95 81 126% 117% [ˈzuʊŋə]






tenero 532 438 1571 1608 44 53 81 54% 65% [ˈtenɐrə] 569 0 1610 0 91 0 105 86% 0 [ˈtenɐr]





F1 F2 Duration v/VF2

GB

Transcr.

DP

MM

LB

AS

FormsF1

ˈσ.CEC] C ˈσ.CEC] #

43/60


Discussion and provisional conclusions

Intrusive vs. Epenthetic vowel (Hall 2006)

Intrusive vowel




• if copied, the intervening

consonant is a sonorant or a

guttural

• generally occurs in heterorganic

clusters





Epenthetic vowel


nearby vowel


• if copied, no restrictions on the

intervening consonants


speech rate


44/60



Intrusive vs. Epenthetic vowel

Intrusive vowel








Carr. I Pontr. I Pontr. E

✓ ✓ X

✓ ✓ X

✓ ✓ X

45/60




Epenthetic vowel


nearby vowel



speech rate


Carr. I Pontr. I Pontr. E

X X ✓

X X ✓

X X ✓

X X ✓

46/60


ˈσ.CCv] WF if SC

• Carrarese I WM if SC

ˈσ.Cv1Cv2] WM/WF if SC

R/I I E WM/(WF) if SC

[ə] [ə] [a/u] ˈσ.CCv] WF if SC

I

ˈσ.Cv1Cv2] WF if SC

• Pontremolese

ˈσ.CCv] WM if SC

E

ˈσ.Cv1Cv2] WM if SC



47/60



Diachrony

Diatopic variation ≈ Diachronic variation

• “This connection between rule generalization and geographical space

arises because sound change originates in a focal area […], from which it

propagates outwards in line with Schmidt’s […] wave theory.”

(Bermúdez-Otero 2014)

• “the faint vowel as the ‘joining link’ between the unstressed vowel

persistence of the Tuscan dialect and the constant deletion of the Emilian

one” (Luciani 1999: 82) … and between the epenthetic vowel of the latter

and the lack of it of the former

• Today’s phonology is yesterday’s phonetics (adapted from Givón 1971:

“today’s morphology is yesterday’s syntax”)

48/60



Diachrony - Phonologization

Diatopic variation ≈ Diachronic variation ≈ Phonologization

Carrarese I: • Optional and unpredictable

(ˈσ.CCə]/ˈσ.CC]/ˈσ.CIC]/*ˈσ.CICə])

• extreme short (and variable) duration

• speech rate dependent

Pontremolese E: • obligatory

• less short and variable

• non-speech rate dependent

The Carr. I represents the diachronically preceding stage of the Pontr. E:

the articulation/perception-conditioned (Hall 2006; Hyman 2013) ə-like

release (Carr.) has (diachronically) been phonologized into /a u/ (Pontr.)

49/60



Phonologization

Phonetics Phonology gradient categorical

continuous discrete

quantitative qualitative

physical symbolic

“[…] what begins as an intrinsic byproduct of something, predicted by

universal phonetic principles, ends up unpredictable, and hence,

extrinsic.” (Hyman 1976:408)

universal phonetics language-specific phonetics phonology

(“automatic”) (“speaker-controlled”) (“structured”)

C release Carr. I [ə] Pontr. E /a u/

50/60



Phonologization




Bermúdez-Otero (2014):

I: “[…] a phonetic phenomenon that is at first exhaustively determined by

extragrammatical factors (physics and physiology) becomes even more deeply

embedded in the grammar of a language, first as a language-specific gradient

process of phonetic implementation”

II: “[…] later as a categorical phonological rule […] until it eventually escapes

phonological control altogether […]

“This occurs when a listener/learner misinterprets the effects of a purely physical or physiological

phenomenon as being under the control of speakers’ grammars, and so adjusts her phonetic

implementation rules accordingly.”

51/60



Phonologization




Hall (2006):

• “[t]he ‘segmentalisation’ (Harms 1976) of intrusive vowels is likely a case of

listener-initiated sound change (Ohala 1981). If intrusive vowels become

acoustically too similar to segmental vowels, speakers may reanalyze them as

segments (Browman & Goldstein 1990a: 318) […] As Hyman (2001: 153)

observes, a phonologized pattern becomes subject to ‘structural or systemic

principles’ that can make it different from its phonetic predecessor […] Once

the vowel has its own gesture, its quality is no longer determined by purely

phonetic considerations, and can be influenced by other characteristics of the

grammar.”

52/60



Phonologization

From Bermúdez-Otero (2014):

Carrarese I Carrarese I (+RV) Pontremolese E

/u/

I vocoid [ə]

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

uE

/u/

/i/

aE

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1


unstressed /a/

/u/

/i/

aE

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

word-medial vocoid [ə]


/u/

/i/

/a/

Vowels

05001.0001.5002.0002.500

F2

200

400

600

800

F1

53/60



Vowel insertion as hypocorrection

Enhancement of the vowel-like release as an Ohalian hypocorrection:

“a process whereby the listener parses a perceptual cue b1 of the context

(segment B) as belonging to segment A. In hypercorrection, a perceptual

cue a2 of a segment A is interpreted as belonging to the context (segment

B)[…] both processes involve perceptual cues that change their association

from one segment to another.” (Hamann 2009)

54/60


/ut/ > [yt] [yt] > /yt/ /yt/ > [yt] [yt] > /ut/

• The speaker ‘automatically’ raises The speaker faithfully realizes the

/u/ F2 (anticipatory coarticulation) phonological high F2 of /y/

• The listener does not compensate The listener interprets the high F2 as

for the ‘automatic’ F2 raising and determined by the coronal context and

interprets it as a phonological feature not as a vowel’s phonological feature

belonging to the vowel

55/60




• STAGE I: /gr/ > [g(ə)r] [g(ə)r] > /gər/ • The speaker ‘automatically’ produces a schwa-like release

• Articulation: low degree of overlap between the two consonant gestures

• Perception: enhancement of the consonat place perception

• The listener interprets this ‘phonetic’ byprodut as being under the

speaker’s control and starts considering the epenthesis as a

phonological process • Alternative: same phonological representation but a change in the phonetic

implementation

• STAGE II: /gər/ > [gɐr] [gɐr] > /gar/ • The speaker colours the schwa

• Coarticulation

• The listener interprets this ‘phonetic’ colouring as being under the

speaker’s control (spreading of the |A| element from /r l/)

56/60





Perception Grammar (Hamann 2009)

Two acquisition stages:

I: “at the age of 6-8 months, the

infants focus on a few, presumably

the most salient, auditory

dimensions […]. Based on the

statistical distributions, the infants

construct language-specific

phonetic categories”

• II: “the infants acquire labels for the

learnt phonetic categories, and are

guided in this by the lexicon. The

lexicon informs the learners about

the abstract categories necessary to

distinguish words in the language.”

57/60



Cue constraints (Hamann 2009)

• “The task of a perception grammar is to map each discrete auditory input to

an abstract phonological category”

• “Cue constraints are not universal and depend both on the perceptual cues

and the phonological categories that are employed in the respective language”

• Cue constraints for /ut/ > /yt/ hypocorrection:

• [low F2] V /back/: low F2 values in the vowel cue back vowels

• [low F2] V /–front/: low F2 values in the vowel cue non-front vowels

• [high F2] VC /–back/: high F2 values in VC transitions cue non-back vowels

• [high F2] VC /cor/: high F2 values in VC transitions cue coronals

58/60



Cue constraints

• “Learners acquire a different weighting because one cue is less reliable than

it used to be for former generations”

Generation x’s grammar: Generation x+1’s grammar:

• [high F2]vc is a cue for t’s /cor/ in generation x, but for both t’s /cor/ and

y’s /-back/ in generation x+1(creation of a new cue constraint)

• /yt/ is the winner candidate (i.e., the correct phonological representation)

because of the high ranking of [high F2]vc /-back/ with respect to [high

F2]vc /cor/

59/60


References Andersen H. (1988). Center and periphery: adoption, diffusion, and spread. In J. Fisiak (ed.), Historical Dialectology, Regional and

Social. Berlin: Mouton de Gruyter, pp. 39–83.

Bertoni G. & Bartoli M. (1925). Breviario di neolinguistica. Modena: Società Tipografica Modenese.

Bermúdez-Otero R. (2014). Amphichronic explanation and the life cycle of phonological processes. In P. Honeybone & J. C.

Salmons (eds), The Oxford handbook of historical phonology. Oxford: Oxford University Press

Blevins J. (2004). Evolutionary phonology: The emergence of sound patterns. Cambridge University Press.

Browman C. P., & Goldstein, L. (1990). Gestural specification using dynamically-defined articulatory structures. Journal of

Phonetics, 18(3), 299-320.

Giannarelli D. (1913). Studi sui dialetti Lunigianesi compresi fra la Magra e l’Appennino Reggiano. Revue de dialectologie romane,

V, pp. 261-311.

Givón T. (1971, April). Historical syntax and synchronic morphology: an archaeologist's field trip. In Chicago Linguistic

Society (Vol. 7, No. 1, pp. 394-415).

Hall N. (2006). Cross-linguistic patterns of vowel intrusion. Phonology, 23, pp. 387–429.

Hamann S. (2009). The learner of a perception grammar as a source of sound change. In P. Boersma & S. Hamann

(Eds.), Phonology in Perception. Berlin: Mouton de Gruyter, 111–149.

Harms R. T. (1976). The segmentalization of Finnish ‘nonrules’. In Texas linguistic forum (Vol. 5, pp. 73-88).

Hyman L. M. (1976). Phonologization. Linguistic studies offered to Joseph Greenberg, 2, 407-418.

Hyman L. M. (2001). The limits of phonetic determinism in phonology:* NC revisited. The role of speech perception in phonology,

141-185.

Hyman L. M. (2013). Enlarging the scope of phonologization". In A.C.L. Yu (Ed.), Origins of Sound Change: Approaches to

Phonologization, 3, Oxford: Oxford University Press, 3-28.

Loporcaro M. (2009). Profilo linguistico dei dialetti italiani. Bari: Laterza.

Loporcaro M. (2011a). Syllable, segment and prosody. In M. Maiden, J. C. Smith & A. Ledgeway (eds.), The Cambridge History of

the Romance Languages, vol. I.: Structures. Cambridge: Cambridge University Press, 50-108.

Luciani L. (1999). Il dialetto Carrarese. Suoni, forme, costrutti, parole. Carrara: Aldus.

Ohala J. 1. 1981." The listener as a source of sound change.". Papers from the Parasession on Language and Behavior, 178-203.

Rohlfs G. (1966). Grammatica storica della lingua italiana e dei suoi dialetti. Fonetica. Turin: Einaudi.

Trudgill P. (1992). Dialect typology and social structure. In E. H. Jahr (ed.), Language contact: theoretical and empirical studies.

Berlin: Mouton de Gruyter, pp.195-212.

60/60


• ?How to formalize this mapping? • CVCV

• Representation: CVCV

• Computation: • Synchrony: Government, Licensing, etc.

• Diachrony: Change in Governing/Licensing power? (Which maybe causes a) change

in the representation? (An increase in the representation richness, the

more universal is the phonology-phonetics mapping: what has to be

learnt is the (details of the ) representation [see Tobias paper])

x/60


• ?How to formalize this mapping? • OT

• Representation: Generally autosegmental (without empty nuclei)?

• Computation:

• Synchrony: Candidates’ Violations computation

• Diachrony: Constraints re-ranking and/or(if the constraints are not universal

(Hamann RH) creation of new constraints (cue constraints, hence

‘phonetically-grounded; to be ranked together with maybe universal

structural constrains (Tobias handout: syllable-based processes

(“Internal GEN”): “Positional pressure is universal and permanently

present: positions do not become strong or weak all of a sudden when

some effect is observed.”))

• “[BiPhon] is used […] it is the only linguistic model at present that includes an explicit

formalization of the phonetics-phonology interface, more specifically the mapping from

auditory form onto phonological form with so-called cue constraints.

x/60

Epenthesis and intrusive vowels in Lunigiana: diatopy, dichrony and phonologization

Documents