A real time comparison of the lenition of intervocalic /d/ in Spanish: Examining the changes in the...

1
A REAL TIME COMPARISON OF THE LENITION OF INTERVOCALIC /d/ IN SPANISH: Examining the changes in the sociolinguistic profile from 1987 to 2004-2010 Manuel D´ ıaz-Campos, Olga Scrivner, Gibran Delgado Indiana University There is hardly anything to hear, but we are hearing it nonetheless (Figueroa, 2014) BACKGROUND Lenition of intervocalic /d/ is one of the most studied phenomenon in the dialectological and sociolinguistic literature dedicated to Spanish. Zamora-Vicente (1970) Intervocalic /d/ deletion can be documented as far as the 17th century. Cedergren (1973) Provides the sociolinguistic profile of intervocalic /d/ in the Spanish of Panama. Elision is favored in informal styles by women, older speakers, and lower socioeconomic participants from rural areas. Navarro Tom´ as (1999:101) /d/ is produced as a fricative in -ado participial forms in formal situations, while in spontaneous speech /d/ is weakened or deleted. BACKGROUND: VENEZUELA Studies analyzing intervocalic /d/ in Venezuelan Spanish include D’Introno and Sosa (1986) and D´ ıaz-Campos and Gradoville (2011). D’Introno and Sosa (1986) Three variants of the intervocalic /d/ variable are: a) retention, b) transition (a weakened or approximant production) and 3) deletion. Retention - formal styles upper and middle class speakers Deletion - the lower socioeconomic class and male speakers GOALS OF THE PRESENT INVESTIGATION 1. Comparing corpora from two points in real time 2. Analyzing the dependent variable as continuous by performing adapted acoustic measurements based on Carrasco, Hualde and Simonet (2012) investigation CORPUS Thirty six speakers from Diachronic Study of the Speech of Caracas 1987 and 2004-2010 were selected for the analysis. The speakers are evenly divided according to sex, socioeconomic level and age. 0.80 0.85 0.90 0.95 1.00 high low mid Intensity Ratio Following Context for D1 0.80 0.85 0.90 0.95 1.00 high low mid Intensity Ratio Preceding Context for D1 0.80 0.85 0.90 0.95 1.00 adj adv noun part pro verb Intensity Ratio Category for D1 Random Forest Analysis Class Sex Period Age 0.00008 0.00012 0.00016 Variable Importance (Social Categories) v2stress Diphthongue1 v1stress Diphthtongue2 PrecedingContext FollowingContext totalDuration 0.00000 0.00005 0.00010 0.00015 Variable Importance (Phonetics) itemLog lexemaLog Category 8.0e-06 1.0e-05 1.2e-05 1.4e-05 1.6e-05 Variable Importance (Frequency/Category) Conditional Tree Analysis FollowingContext p < 0.001 1 {low, mid} high Node 2 (n = 968) 0.8 0.85 0.9 0.95 1 v2stress p = 0.031 3 non-stressed stressed Node 4 (n = 29) 0.8 0.85 0.9 0.95 1 Node 5 (n = 34) 0.8 0.85 0.9 0.95 1 Age p = 0.004 1 {2034, 3554} 55+ Sex p < 0.001 2 m f Period p < 0.001 3 87 2004 Node 4 (n = 172) ! ! ! ! ! ! ! 0.8 0.85 0.9 0.95 1 Age p = 0.013 5 2034 3554 Node 6 (n = 87) ! 0.8 0.85 0.9 0.95 1 Node 7 (n = 80) ! ! ! 0.8 0.85 0.9 0.95 1 Node 8 (n = 347) ! ! ! ! ! ! 0.8 0.85 0.9 0.95 1 Node 9 (n = 345) ! ! ! ! ! ! 0.8 0.85 0.9 0.95 1 lexemaLog p < 0.001 1 1.94 > 1.94 Node 2 (n = 632) 0.8 0.85 0.9 0.95 1 Node 3 (n = 399) 0.8 0.85 0.9 0.95 1 Mixed Eects Models (Speakers and Tokens) Variable 1987 Value p-value Age: 20-34 0.014 0.0002 Age: 35-54 0.017 0.0000 Following Context: low 0.016 0.01 Following Context: mid 0.013 0.045 Frequency lexema 0.005 0.003 Variable 2004/10 Value p-value Age: 35-54 -0.008 0.03 Following Context: low 0.027 0.0001 Following Context: mid 0.022 0.0009 Sex: m 0.008 0.006 Category participle 0.010 0.018 Dependent Variable: Intensity Ratio Independent Variables the lowest intensity point of /d/ the highest intensity point of the preceding vowel Proceeding and following phonetic context, stress, grammatical category, frequency, sex, socioeconomic level and age. INTENSITY RATIO 0 5 10 15 0.80 0.85 0.90 0.95 1.00 Density Intensity Difference (CV) The intensity ratio equation provides a value between 1, a more vowel like production, and 0, a more stop like production. Since intervocalic /d/ tends to be produced as an approximant in intervocalic position a 0 value is not expected. The lowest ratio was 0.85. DISCUSSION 1. The phonetic context maintains its eects in both periods suggesting that coarticulatory factors keep playing a role in explaining lenition of intervocalic /d/. 2. The findings reveal that frequency of the lemma was selected as significant in the 1987 corpus. However, it was not selected for the 2004-2010 corpus. These results suggest that the change has spread extensively in all favoring contexts. 3. The 2004-2010 data show that past participles favor lenition of intervocalic /d/ which maybe related to a type frequency eect. 4. Regarding social factors, age plays an interesting role possibly indicating the profile of a change advancing in 1987 and a more stable profile in 2004/2010 corpus. The fact that men favor this phenomenon in 2004-2010 maybe related to its social meaning as a part of the vernacular. CONCLUSIONS While this phenomenon has been documented since the 17th century and it can be considered a vernacular variant in the speech community, in the most recent period, the sociolinguistic profile has changed. Variation is widespread across all age groups and socioeconomic classes. However, it seems that lenited variants are associated with male speech. References Carrasco, Patricio, Jos´ e I. Hualde and Miquel Simonet. 2012. Dialectal dierences in Spanish voiced obstruent allophony: Costa Rican versus Iberian Spanish. Phonetica, 69, 149-179 Cerdergren, Henrietta. 1973. The interplay of social and linguistic factors in Panama. Doctoral dissertation. Cornell University D’Introno, Francesco, and Juan Manuel Sosa. 1986. La elisi ´ on de la /d/ en el espa ˜ nol de Caracas: Aspectos socioling ¨ ısticos e implicaciones te´ oricas. Estudios sobre la fonolog´ ıa del espa ˜ nol del Caribe, ed. by Rafael N ´ nez Cedeno, Iraset P ´ aez, and Jorge Guitart, 135-163. Caracas: La Casa de Bello ıaz-Campos, Manuel, and Michael Gradoville. 2011. An Analysis of Frequency as a Factor Contributing to the Diusion of Variable Phenomena: Evidence from Spanish Data. In Selected Proceedings of the 2009 Hispanic Linguistics Symposium, ed. Ortiz-Lopez, Luis. Somerville, MA Hualde, Jos´ e I., Miquel Simonet and Mariana Nadeu. 2011. Consonant lenition and phonological recategorization. Lab. Phonology. 2: 301-329 The authors would like to thank Paola Bentivoglio and Kristel Guirado (Universidad Central de Venezuela) and Stephanie Dickinson (Indiana University) [email protected], [email protected], [email protected]

Transcript of A real time comparison of the lenition of intervocalic /d/ in Spanish: Examining the changes in the...

Page 1: A real time comparison of the lenition of intervocalic /d/ in Spanish: Examining the changes in the sociolinguistic profile from 1987 to 2004-2010

A REAL TIME COMPARISON OF THE LENITION OF INTERVOCALIC /d/ IN SPANISH:Examining the changes in the sociolinguistic profile from 1987 to 2004-2010

Manuel Dıaz-Campos, Olga Scrivner, Gibran DelgadoIndiana University

There is hardly anything to hear, but we are hearing it nonetheless (Figueroa, 2014)

BACKGROUNDLenition of intervocalic /d/ is one of the most studied phenomenon inthe dialectological and sociolinguistic literature dedicated to Spanish.

Zamora-Vicente (1970)Intervocalic /d/ deletion can be documented as far as the 17th century.

Cedergren (1973)Provides the sociolinguistic profile of intervocalic /d/ in the Spanish ofPanama. Elision is favored in informal styles by women, olderspeakers, and lower socioeconomic participants from rural areas.

Navarro Tomas (1999:101)/d/ is produced as a fricative in -ado participial forms in formalsituations, while in spontaneous speech /d/ is weakened or deleted.

BACKGROUND: VENEZUELA

Studies analyzing intervocalic /d/ in Venezuelan Spanish includeD’Introno and Sosa (1986) and Dıaz-Campos and Gradoville (2011).

D’Introno and Sosa (1986)Three variants of the intervocalic /d/ variable are: a) retention, b)transition (a weakened or approximant production) and 3) deletion.

Retention - formal styles upper and middle class speakers

Deletion - the lower socioeconomic class and male speakers

GOALS OF THE PRESENT INVESTIGATION

1. Comparing corpora from two points in real time

2. Analyzing the dependent variable as continuous by performingadapted acoustic measurements based on Carrasco, Hualde andSimonet (2012) investigation

CORPUSThirty six speakers from Diachronic Study of the Speech of Caracas 1987and 2004-2010 were selected for the analysis. The speakers are evenlydivided according to sex, socioeconomic level and age.

●●

●●●

●●

●●●

●●

0.80

0.85

0.90

0.95

1.00

high low mid

Inte

nsity

Rat

io

Following Context for D1

●●

●●

●●●

●●

0.80

0.85

0.90

0.95

1.00

high low mid

Inte

nsity

Rat

io

Preceding Context for D1

●●

●●

●●

0.80

0.85

0.90

0.95

1.00

adj adv noun part pro verb

Inte

nsity

Rat

io

Category for D1

Random Forest Analysis

Class

Sex

Period

Age

0.00008 0.00012 0.00016

Variable Importance (Social Categories)

v2stress

Diphthongue1

v1stress

Diphthtongue2

PrecedingContext

FollowingContext

totalDuration

0.00000 0.00005 0.00010 0.00015

Variable Importance (Phonetics)

itemLog

lexemaLog

Category

8.0e−06 1.0e−05 1.2e−05 1.4e−05 1.6e−05

Variable Importance (Frequency/Category)

Conditional Tree Analysis

FollowingContextp < 0.001

1

{low, mid} high

Node 2 (n = 968)

●●●●●

●●

●●

0.8

0.85

0.9

0.95

1

v2stressp = 0.031

3

non−stressed stressed

Node 4 (n = 29)

0.8

0.85

0.9

0.95

1

Node 5 (n = 34)

0.8

0.85

0.9

0.95

1

Agep = 0.004

1

{20−34, 35−54} 55+

Sexp < 0.001

2

m f

Periodp < 0.001

3

87 2004

Node 4 (n = 172)

!

!!!!!!

0.80.850.9

0.951

Agep = 0.013

5

20−34 35−54

Node 6 (n = 87)

!

0.80.850.9

0.951

Node 7 (n = 80)

!!

!0.80.850.9

0.951

Node 8 (n = 347)

!

!!

!!

!0.8

0.850.9

0.951

Node 9 (n = 345)

!

!

!!!

!

0.80.850.9

0.951

lexemaLogp < 0.001

1

≤ 1.94 > 1.94

Node 2 (n = 632)

●●

0.8

0.85

0.9

0.95

1

Node 3 (n = 399)

0.8

0.85

0.9

0.95

1

Mixed Effects Models (Speakers and Tokens)

Variable 1987 Value p-valueAge: 20-34 0.014 0.0002Age: 35-54 0.017 0.0000Following Context: low 0.016 0.01Following Context: mid 0.013 0.045Frequency lexema 0.005 0.003

Variable 2004/10 Value p-valueAge: 35-54 -0.008 0.03Following Context: low 0.027 0.0001Following Context: mid 0.022 0.0009Sex: m 0.008 0.006Category participle 0.010 0.018

Dependent Variable: Intensity Ratio Independent Variables

the lowest intensity point of /d/

the highest intensity point of the preceding vowel

Proceeding and following phonetic context,stress, grammatical category, frequency, sex,socioeconomic level and age.

INTENSITY RATIO

0

5

10

15

0.80 0.85 0.90 0.95 1.00

Den

sity

Intensity Difference (CV)The intensity ratio equationprovides a value between 1, a more vowel likeproduction, and 0, a more stop like production.

Since intervocalic /d/ tends to be producedas an approximant in intervocalic position a 0value is not expected. The lowest ratio was 0.85.

DISCUSSION

1. The phonetic context maintains its effects in both periods suggestingthat coarticulatory factors keep playing a role in explaining lenitionof intervocalic /d/.

2. The findings reveal that frequency of the lemma was selected assignificant in the 1987 corpus. However, it was not selected for the2004-2010 corpus. These results suggest that the change has spreadextensively in all favoring contexts.

3. The 2004-2010 data show that past participles favor lenition ofintervocalic /d/ which maybe related to a type frequency effect.

4. Regarding social factors, age plays an interesting role possiblyindicating the profile of a change advancing in 1987 and a morestable profile in 2004/2010 corpus. The fact that men favor thisphenomenon in 2004-2010 maybe related to its social meaning asa part of the vernacular.

CONCLUSIONSWhile this phenomenon has been documented since the 17th centuryand it can be considered a vernacular variant in the speech community,in the most recent period, the sociolinguistic profile has changed.

Variation is widespread across all age groups and socioeconomic classes.However, it seems that lenited variants are associated with male speech.

ReferencesCarrasco, Patricio, Jose I. Hualde and Miquel Simonet. 2012. Dialectal differences in Spanish voiced obstruent allophony: Costa Rican versusIberian Spanish. Phonetica, 69, 149-179

Cerdergren, Henrietta. 1973. The interplay of social and linguistic factors in Panama. Doctoral dissertation. Cornell University

D’Introno, Francesco, and Juan Manuel Sosa. 1986. La elision de la /d/ en el espanol de Caracas: Aspectos sociolinguısticos e implicacionesteoricas. Estudios sobre la fonologıa del espanol del Caribe, ed. by Rafael Nunez Cedeno, Iraset Paez, and Jorge Guitart, 135-163. Caracas: LaCasa de Bello

Dıaz-Campos, Manuel, and Michael Gradoville. 2011. An Analysis of Frequency as a Factor Contributing to the Diffusion of Variable Phenomena:Evidence from Spanish Data. In Selected Proceedings of the 2009 Hispanic Linguistics Symposium, ed. Ortiz-Lopez, Luis. Somerville, MA

Hualde, Jose I., Miquel Simonet and Mariana Nadeu. 2011. Consonant lenition and phonological recategorization. Lab. Phonology. 2: 301-329

The authors would like to thank Paola Bentivoglio and Kristel Guirado (Universidad Central de Venezuela) and Stephanie Dickinson (Indiana University) [email protected], [email protected], [email protected]