Original Paper

Phonetica 2015;72:121–137 Received: July 1, 2014

DOI: 10.1159/000439599 Accepted after revision: August 21, 2014
Published online: December 22, 2015

From Reduction to Apocope: Final Poststressed

Vowel Devoicing in Brazilian Portuguese

Francisco Menesesa Eleonora Albanoa, b

a
Institute of Language Studies, University of Campinas, Campinas, and
b
CNPq (Brazilian National Research Council), Brasilia, Brazil

Abstract
This is a study of final poststressed vowel devoicing following /s/ in Brazilian
Portuguese. We contradict the literature describing it as deletion by arguing, first,
that the vowel is not deleted, but overlapped and devoiced by the /s/, and, second,
that gradient reduction with devoicing may lead to apocope diachronically. The
following results support our view: (1) partially devoiced vowels are centralized;
(2) centralization is inversely proportional to duration; (3) total devoicing is accom-
panied by lowering of the /s/ centroid; (4) the /s/ noise seems to be lengthened
when the vowel is totally devoiced; (5) aerodynamic tests reveal that lengthened
/s/ has a final vowel-like portion, too short to be voiced; (6) lengthened /s/ favors
vowel recovery in perceptual tests. This seems to be a likely path from reduction
to devoicing to listener-based apocope.
© 2015 S. Karger AG, Basel

1 Introduction

This paper analyzes an allophonic process that has been commonly described as
deletion of poststressed vowels in Brazilian Portuguese (henceforth BP).
Portuguese went through several sound changes throughout its history. Vowel apo-
cope is among the most recurrent, e.g. amare > amar; legale > leal, mense > mês. It
is in fact pervasive in Romance languages, given its Vulgar Latin origins. Deletion
tends to apply primarily to poststressed syllables in languages that have maintained the
asymmetric stress contour of Vulgar Latin (a reinterpretation of the Latin stress rule in
which intensity falls abruptly after stress). Later it spread to prestressed syllables – as
syncope (Taylor, 1994; Wheeler, 2007; Cunha, 2015, this volume) – in languages tend-
ing to stress timing and left-headedness, such as Catalan and European Portuguese
(Ramus et al., 2003). However, the conservative stress contours of present-day BP,
Galician and Spanish are more likely to favor apocope.
Recently, several studies have claimed that final vowels tend to be deleted in
weak positions in BP (Pagel, 1993; Viegas and Oliveira, 2008; Rolo and Mota, 2012).
However, studies of similar cases in other languages such as English, Turkish, French
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

© 2015 S. Karger AG, Basel Francisco Meneses

0031–8388/15/0723–0121 LAFAPE – Laboratório de Fonética e Psicolinguística
$39.50/0 Rua Sérgio Buarque de Holanda 571
E-Mail karger@karger.com Campinas, São Paulo 13083-859 (Brazil)
Downloaded by:

www.karger.com/pho E-Mail chico.meneses@gmail.com

and Japanese have argued convincingly that the phenomenon is not apocope, but
devoicing (Jannedy, 1995; Tsuchida, 1997; Smith, 2003; Davidson, 2006). Despite cer-
tain speculations based on historical records (Lemle, 1966; Viaro, 2005), only recently
was devoicing recognized as such in BP (Meneses, 2012; Meneses and Albano,
2013). There are also reports of a related, more advanced sound change in European
Portuguese (e.g. Fernandes, 2007; see also Cunha, 2015, this volume).
In view of these facts, a first aim of this study is to show that there is a syn-
chronic variation process in BP that ranges from reduction to total devoicing of vow-
els in a final poststressed position. Such variation can lead to ambiguities that may
result diachronically in listener-based apocope (Ohala, 1981). A second aim of this
study is to argue for the interpretation of the steps of this change in light of artic-
ulatory phonology (henceforth AP). AP explains several phonological phenomena
by means of 3 dynamical processes: gestural overlap, gestural blending and gesture
magnitude reduction (Saltzman and Munhall, 1989; Browman and Goldstein, 1990;
Romero, 1996; Iskarous et al., 2012). Extreme reduction leading to devoicing and,
eventually, apocope is interpreted here as a prosodically driven decrease in gestural
magnitude accompanied by an increase in gestural overlap between the vowel and
the preceding consonant.

2 Background

2.1 The Path from Reduction to Devoicing

Vowel reduction is a common phenomenon across languages and has been well
documented since Lindblom (1963). It has been discussed mainly in terms of percep-
tion, since listeners have strategies to recover the identity of vowels in spite of their
acoustic alterations (Van Bergem 1995; Beckman et al., 1992). According to Lindblom
(1963), such alterations include centralization of F1/F2 and shorter vowel duration.
Reduction is also related to stress, speech rate and linguistic as well as pragmatic con-
text, ranging from hypo- to hyperarticulation (for discussion of vowel reduction in
Spanish/English bilinguals, see Fabra, 2015, this volume).
While the mechanisms of vowel reduction seem to be well understood, devoic-
ing still needs clarification, since it obliterates most default properties of vowels such
as voicing, periodic waveform and regular formant structure (Han, 1962; Tsuchida,
1997). When preceded by a fricative, devoiced vowels look like an acoustic continua-
tion of the preceding consonant (Kondo, 1997, 2005). Kondo (1997) considers devoic-
ing more than a simple loss of voice, and goes on to interpret it as one of the steps
in what she terms ‘vowel weakening’, namely reduction, devoicing and, eventually,
deletion.
The literature on vowel devoicing has so far produced a single, very preliminary
study within the AP framework (Chitoran and Iskarous, 2008). Based on acoustic data,
the authors argue that devoicing in Lezgi results from gestural changes that lead to
overlap between adjacent consonantal and vocalic articulatory gestures. They explain
devoicing in terms of alterations in the magnitude and timing of such gestures.
Despite having made advances in the description of vowel reduction and devoic-
ing as independent phenomena, the current literature has not gone beyond Kondo’s
(2005) initial proposal in investigating their relationship. Neither has it tried to unify
their theoretical treatment. We will attempt to fill this gap.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

122 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
 2.2 Synchronic Variation of Vowels in BP
The best clues to the relationship between reduction and devoicing lie in historical
grammars, which record instability of vowels in environments where they later disap-
pear. To understand how these may have affected BP, two considerations are necessary.
The first concerns the reduction of the vowel inventory depending on stress.
According to Câmara Jr. (1953, 1970), BP vowels undergo neutralization as follows:
there are 7 vowels, /a, ɛ, e, i, ɔ, o, u/, in the stressed position, 5 vowels, /a, e, i, o, u/,
in the prestressed position, 4 vowels, /a, i, o, u/, in the medial poststressed position
and only 3 vowels, /a, i, u/, in the final poststressed position. This contributes to the
asymmetric nature of the BP stress contour, which starts high, rises smoothly up to the
stressed vowel and then falls abruptly.
The second point to consider is the extreme weakness of the poststressed posi-
tion, which, according to Câmara Jr. (1970), carries zero stress in a scale from 0 to 3.
Consequently, this position is the most susceptible to radical reduction and devoic-
ing of the kind examined below (Aquino, 1997; Ferreira, 2007; Souza, 2012; Dubiela,
2013).
This is why the final poststressed position was the focus of Meneses’s (2012)
study, which aimed at uncovering the gradient side of vowel ‘apocope’. The argument
is that these vowels are not completely deleted, but just partially or totally devoiced.
Here we will depart from Meneses’s findings to show that vowel reduction in post-
stressed position constitutes a continuum that is likely to be related to the steps in
vowel deletion proposed by Kondo (2005). However, unlike Kondo, and in line with
Chitoran and Iskarous (2008), we will try to treat this process as a synchronic gradient
phenomenon and give it a dynamical theoretical interpretation.

2.3 Sound Change in Progress

There are at least 3 perspectives that can contribute to the understanding of the
steps involved in the changes that lead from reduction to devoicing and apocope. The
first is variationist sociolinguistics, which has established that sound change emerges
from synchronic variability (Labov, 1972). It conceives of change as gradual, evolving
from generation to generation, in response to sociolinguistic factors such as age, educa-
tional level, socioeconomic status, etc.
The second perspective sees change as more abrupt. It claims that the perception
of a sound different from the intended one can arise when a listener cannot compensate
for the effects of context (Ohala, 1981, 1993). Thus, any misapprehension or mispars-
ing can be the beginning of a sound change. Diachronic variation also emerges from
synchronic variation in this light.
The third view also considers change gradual but assumes that it can occur not
only across speakers and generations, but also within the same speaker (Harrington et
al., 1995; Harrington 2012). Both of Harrington’s studies are actually a development
of the gestural view of sound change proposed by Browman and Goldstein (1989,
1990), which claims that gestural overlap plays a key role in the variation and sta-
bilization of phonological patterns. Harrington and colleagues (1995, 2012) focus
on a form of overlap between two gestures sharing the same articulator, known as
blending of articulatory movements. Blending is a form of extreme overlap1 and can

1
This is different from simple in-phase coordination, where the vowel reaches its target after the consonant
offset. In blending, the vowel reaches its target before the consonant offset.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 123

Portuguese DOI: 10.1159/000439599
Downloaded by:
Table 1. Target words used in experiment 1

Target vowel Target word

caça ‘hunting’ [‘ka.sa]

/a/ massa ‘mass’ [‘ma.sa]
taça ‘cup’ [‘ta.sa]
lance ‘throw’ [‘lã.si]
/i/ passe ‘pass’ [‘pa.si]
tosse ‘cough’ [‘tɔ.si]
aço ‘steel’ [‘a.su]
/u/ baço ‘spleen’ [‘ba.su]
maço ‘pack’ [‘ma.su]

thus contribute to trigger or consolidate sound change, inasmuch as it tends to cause

masking of one segment by the other. According to Zsiga (2000) and Iskarous et al.
(2012), extensive overlap between segments as different as an obstruent and a vowel
(or an approximant) always leads to blending due to the need to adjust the constriction
degree. Here the case for blending is hard to make, in the absence of direct evidence
of smoothing of articulatory trajectories. We will thus call the phenomenon involved
‘extreme overlap’. In addition, we will explore the idea that ‘auditory error’ may facil-
itate sound change.
In order to shed light on devoicing in BP, we will first present, sum up and inter-
pret our acoustic findings in terms of the gestural mechanism of vowel devoicing.
Then, we will try to demonstrate that there is a continuum of vowel reduction mani-
fested synchronically as extreme vowel centralization and partial or total overlap with
the preceding consonant, which results in extremely short, partially or fully devoiced,
fricative vowels. This variation may eventually constitute the raw material for subse-
quent stages of the sound change resulting in apocope. While the early stages of such
change tend to be related to the timing of production, its last stage is often due to a
perceptual reinterpretation of the acoustic signal (Ohala, 1981).

3 Experiment 1

3.1 Speech Material and Participants

To explore the nature of poststressed vowels in BP, we used [´(C)V.CV]
sequences. Ten target words were chosen to include 3 different vowels, namely /a,
i, u/, preceded by /s/2 as seen in table 1. The carrier sentence was designed with the
same structure and number of syllables for each target word, as for instance: ‘O X
parecia perigoso.’ Each speaker read the sentences 5 times. In order to compare the
vowel space area (VSA), the same words were recorded in a voiced environment, as
‘Disse X baixinho’.
The speakers were 6 females (S11…S613) from Vitória da Conquista, a town in
the state of Bahia where final /s/ does not palatalize as in other northeastern dialects.
2
The voiceless fricative /s/ is the only voiceless consonant occurring frequently in coda position in BP.
3
Subscripts refer to the experiment number.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

124 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
Gender was kept constant to facilitate normalization. Age range (25–40 years old),
educational level (undergraduate students) and socioeconomic status (middle class)
were also kept constant.

3.2 Measurements
Several acoustic properties were measured in target vowels. These measurements
aimed at relating reduction to devoicing. The measurements used in experiment 1
were: /s/ noise centroid, and the durations of the target syllable, noise and vowel. In
order to compare contexts favoring and disfavoring devoicing, VSAs were calculated
as an index of vowel dispersion. The correlation between VSA and vowel duration was
also computed.

3.3 Acoustic Analysis

Devoicing was classified as partial when the vowel overlapped by the consonant
was mostly aperiodic but nevertheless had a short periodic portion with few glottal
pulses (average = 4 pulses) and reduced duration (average = 20 ms). Devoicing was
classified as total when the vowel was completely aperiodic with no glottal pulse and
indefinite formant configuration. Note that these are not equally amenable to acoustic
duration measurements. In total devoicing, noise duration accounts for both the conso-
nant and the vowel of the intended syllable, since they are totally overlapped.
In the spectral analysis, measurements were performed automatically by a
PRAAT script designed to: (i) generate a fast Fourier transform window; (ii) calculate
the length of the labeled segment and define its center; (iii) estimate the first spec-
tral moment of the fricative. As overlap between vowel and consonant tends to affect
centroid measurements and such effects are often located in relatively high spectral
regions, measurements were performed using a filter above 4 kHz. This band was
chosen because alveolar fricatives generally have very little energy below this point
(Vaissière, 2009).
VSA evaluates the relative distances of a speaker’s vowels. It is used to quantify
and represent vowel centralization (Kent and Kim, 2003). VSA values were obtained
by the formula: VSA = ABS [(F1i × [F2a – F2u] + F1a × [F2u – F2i] + F1u × [F2i –
F2a])/2]. These spaces were estimated by triangulation rules, according to plane geom-
etry. Centralization is indicated by low VSA values.

3.4 Statistical Analysis

Paired t tests were used to compare fricative centroid values. The Pearson coef-
ficient was used to correlate VSA to vowel duration. Duration differences were evalu-
ated by the Kruskal-Wallis test due to deviation from normality. Duration data were
normalized by the Z score transformation.

3.5 Results
3.5.1 General Findings
In general, our results confirm the acoustic patterns described by the literature
(Jannedy, 1995; Tsuchida, 1997). High vowels lose voicing (sometimes incompletely),
periodicity and formant structure, as shown in figure 1. Note below that, as pointed out
by Han (1962) and Kondo (2005), the formant pattern is not clear in the case of devoic-
ing (fig. 1c) as compared with partial devoicing (fig. 1b) and voicing (fig. 1a), which
exhibits well-defined boundaries between the vowel and the fricative.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 125

Portuguese DOI: 10.1159/000439599
Downloaded by:
 0.04181 0.04971 0.07477

0 0 0

–0.06268 –0.05252 –0.08026

0 0.1189 0 0.1017 0 0.1078
5,000 5,000 5,000
Frequency (Hz)

Frequency (Hz)

Frequency (Hz)
0 0 0
0 0.1189 0 0.1017 0 0.1078
a Time (s) b Time (s) c Time (s)

Fig. 1. Examples of voiced (a), partially devoiced (b) and totally devoiced (c) syllables /si/, respec-
tively, extracted from the word ‘lance’ [’lã.si].

F2 F2
3,000 2,500 2,000 1,500 1,000 500 0 3,000 2,500 2,000 1,500 1,000 500 0
0 0
100 100
200 200
300 300
400 400
500 500
F1

F1
600 600
[a] 700 700
[i] 800 800
[u] 900 900
1,000 1,000
a b

Fig. 2. Vowel space areas of vowels followed by voiceless (a) and voiced consonants (b).

The proportions of devoiced vowels of different heights are also in line with the
literature (Tsuchida, 1997). Percentages of totally devoiced vowels are 46% for /u/,
42% for /i/ and 12% for /a/. Percentages of partial devoicing are also higher for /u/
(45%) and /i/ (38%) compared to /a/ (17%). As mentioned above, this kind of variation,
recurrent across speakers, is the raw material for sound change in progress.

3.5.2 Extreme Vowel Centralization

It has already been mentioned that vowel reduction is favored in poststressed posi-
tion in BP and other languages. Our results show that reduction is even greater than
expected in contexts that trigger devoicing. Figure 2 shows the vowel space areas for
final unstressed vowels preceded by /s/ in two conditions: followed by voiced conso-
nants and followed by voiceless consonants. The space defined by vowels preceding
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

126 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
 80

70
Vowel duration (ms)

60

50

40

30

20
0

00

00

00

00

0
00

00

00

00

00
,0

,0

,0

,0

0,

0,

0,

0,

0,
20

40

60

80

10

12

14

16

18
Fig. 3. Correlation between
VSA
speaker VSAs and their vowel
duration in voiceless contexts.

35 35
/u/
30 30
/a/
Occurrences (%)

Occurrences (%)

25 25 /i/
20 20
15 15
10 10
5 5
0 0
Devoiced Partially Voiced Devoiced Partially Voiced
a devoiced b devoiced

Fig. 4. Occurrence of devoiced vowels, partially devoiced vowels and voiced vowels for speakers 1
(a) and 5 (b), respectively.

voiceless consonants is more compressed than the space defined by vowels preceding
voiced consonants. This strongly suggests that the vowel reduction is more pronounced
in poststressed syllables followed by voiceless consonants, as seen in figure 2a.
Figure 3 shows the correlation between the speakers’ VSAs and the duration of
their voiced and partially devoiced vowels pooled together. There is a clear trend for
more centralized vowels to be shorter than less centralized vowels. The positive cor-
relation [r(24) = 0.53, p = 0.005] indicates that centralization tends to be directly pro-
portional to vowel shortening.
Figure 4 displays differences in devoicing percentage between two extreme speak-
ers, S11 and S51, to help clarify the relationship between centralization and shortening
in figure 3. While S11 (left) devoices practically all high vowels, S51 (right) has a low
percentage of devoicing, producing more voiced or partially devoiced vowels, the vow-
els of S11 tend to be shorter (69 ms average) and more centralized (VSA = 17.02). The
vowels of S51 tend to be longer (91 ms average) and less centralized (VSA = 103.21).
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 127

Portuguese DOI: 10.1159/000439599
Downloaded by:
 7,100

7,000

6,900
Centroid (Hz)

6,800

6,700

6,600
Fig. 5. Spectral moments of
6,500 fricative in syllables ‘without’
Devoiced /i/ Voiced /i/ vowel (devoiced /i/) and with
vowel (voiced /i/).

Thus, we take S11 and S51 as representative of the kind of variation involved in the
steps of sound change under investigation.
It is clear that final unstressed vowels followed by voiceless consonants undergo
even more radical reduction in speakers such as S11.

3.5.3 Centroid Measurements

Figure 5 shows the differences between the /s/ noise centroids preceding devoiced
/i/ and voiced /i/. Such a difference is statistically significant [t(2) = –6.64, p < 0.02].
In words ending in /i/, the difference between the /s/ noise centroids in the pres-
ence and absence of a visible vowel is approximately 200 Hz. The lower centroid
value indicates that the noise contains acoustic cues to the vowel when it is apparently
absent. We should note that the high devoicing rate and the high intra- and interspeaker
variability prevented this analysis to be performed for /u/.

3.5.4 Duration Data

Generally, vowel devoicing is accompanied by temporal compression, due to
heavy coarticulation (Fagyal and Moisset, 1999; Smith, 2003). The following data, on
duration of devoiced final poststressed vowels, support this interpretation and intro-
duce a new finding, namely noise duration in devoicing is longer than elsewhere. The
measurements are summarized in figure 6, exhibiting /s/ noise average durations for all
3 cases: total devoicing, partial devoicing and voicing.
The figure below shows that the /s/ followed by a devoiced vowel is longer than
in the other two cases, for both /i/ (fig. 6b) and /u/ (fig. 6a). The Kruskal-Wallis test
indicates that mean noise durations differ (H = 8.12, d.f. = 2, p < 0.02).
Figure 7 shows that the trend for syllable length is exactly the opposite: those con-
taining devoiced vowels are shorter than those containing partially devoiced or voiced
vowels.
For /i/, again, the Kruskal-Wallis test indicates that average syllable durations
differ (H = 6.95, d.f. = 2, p < 0.03). As to /u/, there was no significant difference,
despite the apparent difference in figure 7b. This is due to small sample size, given
the lower rate of voicing of /u/. Incidentally, this is also observed in stops. In a
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

128 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
 1.5 1.5

1.0 1.0
Duration (Z score)

Duration (Z score)
0.5 0.5

0 0

–0.5 –0.5

–1.0 –1.0

–1.5 –1.5
Devoiced Partially Voiced Devoiced Partially Voiced
a /u/ devoiced /u/ /u/ b /i/ devoiced /i/ /i/

Fig. 6. a, b Duration of /s/ noise with voiced, partially devoiced and totally devoiced vowels.

1.5 1.5

1.0 1.0
Duration (Z score)

Duration (Z score)

0.5 0.5

0 0

–0.5 –0.5

–1.0 –1.0

–1.5 –1.5
Devoiced Partially Voiced Devoiced Partially Voiced
a /si/ devoiced /si/ /si/ b /su/ devoiced /su/ /su/

Fig. 7. a, b Syllable duration with voiced, partially devoiced and devoiced vowel.

preliminary study, we (Albano and Meneses, 2015) tested for devoicing in vowels
preceded by stops. Preliminary results showed the presence of a longer burst in the
absence of a visible vowel. Thus, duration patterns are apparently similar to those
following fricatives4.
Figures 6 and 7 illustrate the tendency for vowels and consonants to lengthen and
shorten depending on the amount of voicing. This is confirmed in figure 8, which plots
consonant durations against vowel durations in cases where devoicing was not com-
plete. The correlation is negative and significant [Pearson r(45) = –0.46, p = 0.001],
confirming that vowel and consonant durations are inversely proportional.
So far, the centroid and duration results point to the existence of some vowel por-
tion in the /s/ noise. Furthermore, the lower centroid and the longer noise seem to result
from the extreme overlap of the vowel and consonant gestures. Such an overlap pre-
cludes voicing, due to the dominance of the fricative gesture. As seen in the waveform
of figure 1, the consonant noise prevails. This is in accordance with the predictions of

4
The analysis of these data is still under way.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 129

Portuguese DOI: 10.1159/000439599
Downloaded by:
 3

2
Vowel duration (Z score)
1

–1

–2

–3
–3 –2 –1 0 1 2 3 Fig. 8. Correlation between
Consonant duration (Z score) normalized vowel and conso-
nantal duration.

Table 2. Triads of words used in experiment 2

/i/ /u/ Without vowel

ace [‘a.si] aço [‘a.su] ás [‘as]

canse [‘kã.si] canso [‘kã.su] cãs [‘kãs]
coce [‘kᴐ.si] coço [‘kᴐ.su] cós [‘kᴐs]
face [‘fa.si] faço [‘fa.su] faz [‘faz]
lance [‘lã.si] lanço [‘lã.su] lãs [‘lãs]
passe [‘pa.si] passo [‘pa.su] paz [‘pas]
trace [‘tɾa.si] traço [‘tɾa.su] traz [‘tɾas]

AP, which attributes more deformability to the least constricted gesture (Browman and
Goldstein, 1987). In an attempt to further clarify these findings, the next experiment
investigates devoiced vowels through aerodynamic data.

4 Experiment 2

4.1 Preliminary Aerodynamic Test

Our results so far seem to point to the existence of vowel correlates in the noise
of devoiced syllables. Aerodynamic experiments can provide more explicit clues to the
dynamics involved in such gestural overlap.

4.2 Speech Material and Participants

We performed an aerodynamic test with 2 participants (S12, S22) of the earlier
experiment. We recorded triads of words as seen in table 2. Again, only female speak-
ers were recorded for comparability with the earlier experiment. Each speaker read
the target sentences 3 times. All target words were recorded with the carrier sentences
‘Digo X paciente’ and ‘Digo X baixinho’.
We used the EVA2 portable workstation, which makes aerodynamic and acoustic
measurements (Ghio and Teston, 2004). Two airflow channels and 2 pressure channels
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

130 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
 Color version available online
1.0 0.6
0.8
Acoustic amplitude

Acoustic amplitude
0.4
0.6
0.4 0.2
0.2
0 0
–0.2 –0.2
–0.4
–0.6 –0.4
–0.8 –0.6
a 88,600 88,800 89,000 89,200 89,400 89,600 ms b 39,400 39,600 39,800 40,000 40,200 40,400 ms
1.2
1.0 1.0
0.8 0.8

Airflow (dm3/s)
0.6
Airflow (dm3/s)

0.4 0.6
0.2 0.4
0 0.2
–0.2 0
–0.4 –0.2
–0.6 –0.4
–0.8
c 88,600 88,800 89,000 89,200 89,400 89,600 ms d 39,400 39,600 39,800 40,000 40,200 40,400 ms

Fig. 9. Waveform (a, b) and oral airflow (c, d) of /s/ with voiced vowel (a, c) and coda /s/ (b, d) for
the target word passe [‘pa.si] and paz [‘pas].

are available to measure oral and nasal airflow along with pharyngeal pressure. Here,
we have just used 1 oral airflow channel.
The simultaneous recording of sound and airflow requires the use of a ‘mouth-
piece’ attached to the microphone’s mechanical stand. A flexible silicone mask is used
to seal off the mouth so as to obtain reliable oral airflow measurements. To this end, the
pressure sensors were calibrated and checked for each speaker. The Phonedit software
was used for data recording and processing.

4.3 Results
Examination of the synchronized airflow and waveform signals shows that the oral
flow of /s + V/ presents two peaks, as shown in figure 9a and c for ‘Digo paz baixinho’.
The first peak refers to the constriction at the beginning of the friction. The valley that
follows refers to the continuous release of the flow during the production of the frica-
tive. The second peak refers to the relaxation of the constriction to produce the vowel.
In coda /s/ (fig. 9b, d), there is only one oral airflow peak related to fricative
production, followed by the constriction release, since the /s/ is in coda (i.e. no vowel
follows). If we compare figures 9 and 10, we see that there is a clear similarity between
voiced vowels and devoiced vowels: the noise occurring with devoiced vowels also
displays, first, a fricative constriction peak and, second, a peak analogous to that of
voiced vowels. Note the contrast with coda /s/ above.
Figure 10 suggests that neither time nor airflow is sufficient for the production of
a canonical voiced vowel when a voiceless consonant follows. Apparently, voicing is
precluded by the dominance of the fricative gesture, as predicted by the extreme over-
lap analysis.

5 Experiment 3

5.1 Identification Test

As seen in section 3.5, noise lengthening occurs when the vowel is fully devoiced.
Our results so far suggest that lengthening is actually due to the presence of a fricative
vowel, i.e. a vowel that is unable to initiate voicing due to extreme gestural overlap.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 131

Portuguese DOI: 10.1159/000439599
Downloaded by:
 Color version available online
0.6

0.4

Acoustic amplitude 0.2

–0.2

–0.4

–0.6
a 43,200 43,400 43,600 43,800 44,000 ms
1.2
1.0
0.8
Airflow (dm3/s)

0.6
0.4
0.2
0
–0.2
–0.4

b 43,200 43,400 43,600 43,800 44,000 ms

Fig. 10. Waveform (a) and oral airflow (b) of devoiced vowel for the target word passe [‘pa.si] in the
carrier sentence ‘Digo passe paciente’.

In an analogous situation involving blending of tongue body gestures, Iskarous et al.

(2012) noted that the constriction degree was always biased toward consonantal values.
By analogy, the present experiment investigates whether /s/ noise lengthening has an
influence on vowel detection in total devoicing.

5.2 Speech Material and Participants

We performed a forced identification test with 6 female listeners (L13…L63) of
dialects exhibiting devoicing. Again, age range (25–40 years old), educational level
(graduate students, in this case), and socioeconomic status (middle class) were kept
constant.
Stimuli were selected through the following steps: (i) the same triads of words, such
as [ˈfas], [ˈfa.si] and [ˈfa.su], from experiment 2 (table 2) were recorded by speakers
with a high devoicing rate; we created a carrier sentence for each target word; (ii) typical
occurrences of devoicing were selected; (iii) the duration of preceding /s/ was measured
and normalized by a Z score; (iv) selected tokens were then classified into 3 categories:
lengthened, having mean duration (i.e. Z score equal to zero) or belonging to coda /s/.

5.3 Procedure
The perception test consisted in presenting 35 stimuli to each listener. Listeners
heard the stimuli 5 times and chose the matching word among 3 possibilities. For
instance, after hearing the word [ˈfa.si] with stretched /s/, they were asked to choose
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

132 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
 80

70

60
Occurrences (%)

50

40

30

20

10

0
Hit Error Hit Error Hit Error
Lengthened /s/ Mean length /s/ Coda /s/

Fig. 11. Percentages of hits and errors in the identification of lengthened /s/, mean length /s/ and final
/s/.

among [ˈfas], [ˈfa.si] or [ˈfa.su]. Then, listeners were asked to rate their responses in a
5-point scale ranging from maximal to minimal confidence. Incorrect vowel identifica-
tion (e.g. [ˈpa.su] for [ˈpa.si]) was counted as a hit, since the major issue was vowel
detection. After identification, listeners indicated the confidence level of their answer
(from 1 to 5), and this was used to weight the scores.

5.4 Statistical Analysis

The Kruskall-Wallis test was used to evaluate the difference between the weighted
scores in the 3 conditions.

5.5 Results
Figure 11 shows the rates of hits and errors for all test stimuli. Lengthened noise
presented more hits, namely 76%. The error rate was relatively low in lengthened
(24%) as compared to average noise (42%).
In spite of the confusion caused by forced choice, coda /s/ has a high hit rate,
as expected. Interspeaker variability analogous to that of experiment 1 can also be
observed in our identification experiment. Here, the extreme cases are listeners L13
and L33: while L13 has only 9% errors for lengthened /s/, L33 has as much as 37%.
Even though L33 presents fewer errors than hits, the difference between L13 and L33
suggests that there are people who are more likely to misperceive extremely over-
lapped /s + V/.
When we consider the weighted scores in figure 12, i.e. those resulting from
weighting hits and errors by confidence level, the Kruskal-Wallis test indicates that
lengthened noise is a decisive cue to vowel detection (H = 6.33 , d.f. = 2, p < 0.04).
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 133

Portuguese DOI: 10.1159/000439599
Downloaded by:
 9

Weighted score 8
*

7
*

*
6
Mean
Mean ± SE
5 *
Mean ± SD
Outliers *
* Extremes
4 Fig. 12. Mean score weighted
Lengthened /s/ Mean length /s/ Codas /s/ by confidence for lengthened
/s/, mean length /s/ and final /s/.

Average /s/ and coda /s/ yielded more errors and less listener confidence in the
forced identification test. In contrast, lengthened /s/ yielded more hits and more listener
confidence. Therefore, lengthened noise seems to carry important cues for the recovery
of the devoiced vowel. As we shall see in the next section, our results are explainable
in terms of relative timing with various degrees of gestural overlap with total devoicing
as the extreme case.

6 Discussion

The results of this work show that the acoustics, the aerodynamics and the per-
ception of radically reduced final vowels in BP contradict the phonological literature
claiming that there is deletion in that position. Following up on Meneses’s findings
(2012), the above data show that reduction of final poststressed vowels in BP varies
from moderate to extreme, implying devoicing as it approaches critical shortness lead-
ing to total devoicing in a voiceless environment.
In summary, final vowel reduction is expressed in five ways in our data, namely:
(1) partially devoiced vowels are highly centralized and shorter if followed by a con-
sonant; (2) when the vowel is apparently absent, the average centroid differentiates
between regular and extreme CV overlap; (3) the /s/ noise lengthens when the vowel is
fully devoiced; (4) lengthened /s/ exhibits a final vowel-like rise in airflow revealed by
aerodynamic tests; (5) lengthened /s/ noise facilitates vowel recovery.
Let us now try to integrate the interpretation of the above facts.
When preceded by /s/ and followed by a voiceless consonant, voiced vowels are
very short and highly centralized, as indicated by the decrease in both vowel duration and
VSA. When the vowel is fully devoiced, the consonant gesture dominates. Vowels tend
to shorten gradually and slide under the preceding /s/. When overlap is so extreme that
it produces a fricative vowel, two acoustic cues point to the presence of a vowel gesture.
First, a lowered /s/ centroid signals the overlapped vowel. Second, lengthened /s/ points
to the extremeness of such overlap in full devoicing (as shown in experiment 1). This
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

134 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
fricative vowel is too short and constricted to be able to initiate voicing. Fricative vow-
els are reported in many languages and are characterized as noisier than a typical vowel
(Feng, 2007). Here, the fricative vowel seems to be a further step in the reduction process.
The aerodynamic data of experiment 2 also provide further insight into extreme
overlap between C and V. The airflow data show that the final portion of the constric-
tion is consistent with the presence of a fricative vowel. This is because the vocal tract
configuration is narrower than in voiced vowels, as there is no time to achieve the
opening necessary for a significant drop in supraglottal pressure. These results fit in
well with the identification test results, i.e. the fricative vowel increases the listeners’
sensitivity to the vowel cues in the noise.
As seen above, vowel devoicing is directly proportional to overlap in BP. Thus,
the greater the overlap between the vowel and consonant gestures, the more the vowel
devoices. In other words, the timing of onset and nucleus gestures in the final unstressed
position determines the amount of voicing. Therefore, as predicted by AP, the temporal
dimension of articulation appears to be a crucial component of vowel devoicing in both
synchronic variation and sound change.

7 Conclusion

Our results indicate that the reduction of final unstressed vowels in BP is per-
vasive but gradual, ranging from shortening with partial devoicing to full devoicing.
This variability appears to result from differences in gestural coordination and different
degrees of overlap between consonant and vowel gestures.
Such a gradience is more consistent with extreme C and V overlap than with vowel
deletion. However, as the fricative vowel is extremely short, listeners can be misled by
the weak vowel cues in the acoustic signal so that they fail to recover such an overlap.
If, eventually, overlap with total devoicing spreads diachronically in the population,
apocope can be gradually assumed by more and more listeners, since they have vari-
able degrees of sensitivity to the vowel cues in the noise. Thus, devoiced vowels, i.e.
completely overlapped fricative vowels, can trigger listener-based apocope in the long
run, along the lines advocated by Ohala (1981).

Acknowledgments

This research was supported by the Fundação de Amparo à Pesquisa do Estado de São Paulo,
grant No. 2010/04902-0. Support from the Conselho Nacional de Desenvolvimento Científico e
Tecnológico (CNPq), grant No. 311154/2009-3, is also acknowledged. We also thank Didier Demolin,
Khalil Iskarous, the audience of PAPI 2013, Marina Vigário, Rachel Walker, two anonymous review-
ers and our colleagues from Unicamp’s Laboratório de Fonética e Psicolinguística (LAFAPE) for use-
ful comments and suggestions. Special thanks are due to our participants.

References
Albano E, Meneses F (2015): Novas luzes sobre a Dinâmica Sincrônica e Diacrônica do Desvozeamento Vocálico
(oral presentation). Cadernos de resumos do X Congresso Internacional da Abralin. Bélem, Pará, pp 59–60.
Aquino P (1997): O papel das vogais reduzidas pós-tônicas na construção de um sistema de síntese concatenativa
para o português do Brasil; master degree thesis, Unicamp, Campinas.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 135

Portuguese DOI: 10.1159/000439599
Downloaded by:
Beckman M, De Jong K, Jun S-A, Lee S-H (1992): The interaction of coarticulation and prosody in sound change.
Lang Speech 35:45–58.
Browman CP, Goldstein L (1987): Tiers in articulatory phonology, with some implications for casual speech:
Haskins Laboratories Status Report on Speech Research, SR-92; in Kingston J, Beckman ME (eds): Papers in
Laboratory Phonology I: Between the Grammar and the Physics of Speech. Cambridge, Cambridge University
Press, pp 1–30.
Browman C, Goldstein L (1989): Articulatory gestures as phonological units. Phonology 6:201–251.
Browman C, Goldstein L (1990): Tiers in articulatory phonology, with some implications for casual speech; in
Kingston J, Beckman M (eds): Papers in Laboratory Phonology: Between the Grammar and the Physics.
Cambridge, Cambridge University Press.
Câmara JM Jr (1953): Para o estudo da Fonêmica Portuguesa. Rio de Janeiro, Organização Simões.
Câmara JM Jr (1970): Estrutura da língua portuguesa, ed 15. Petrópolis, Vozes.
Chitoran I, Iskarous K (2008): Acoustic evidence for high vowel devoicing in Lezgi. Proceedings of ISSP 8,
Strasbourg, pp 93–96.
Cunha C (2015): Portuguese lexical clusters and CVC sequences in speech perception and production. Phonetica
72:138–161.
Davidson L (2006): Schwa elision in fast speech: segmental deletion or gestural overlap? Phonetica 63:79–112.
Dubiela MR (2013): A vogal /e/ átona final em falantes curitibanos. Monografia Linguística. Curitiba, UFPR.
Fabra LR (2015): Can nonnative speakers reduce English vowels in a native-like fashion? Evidence from L1-Spanish
L2-English bilinguals. Phonetica 72:162–181.
Fagyal Z, Moisset C (1999): Sound change and articulatory release: where and why are high vowels devoiced in
Parisian French? Proceedings of the XIVth International Congress of Phonetic Sciences, Berkeley, University
of California, pp 309–312.
Feng L (2007): The articulatory and acoustic study of fricative vowels in Suzhou Chinese. 16th ICPhS XVI,
Saarbrücken.
Fernandes AC (2007): Apagamento de vogais átonas em trissílabos proparoxítonos: um contributo para a compreen-
são da supressão vocálica em português europeu; master degree thesis, Universidade do Porto.
Ferreira APP (2007): Pet or Pety? Diferenças entre palavras CVC e CVCV do inglês por aprendizes brasileiros: uma
análise acústica; master degree thesis, Curitiba.
Ghio A, Teston B (2004): Evaluation of the acoustic and aerodynamic constraints of a pneumotachograph for speech
and voice studies. Proceedings of International Conference on Voice Physiology and Biomechanics, Marseille,
Université de la Méditerranée, pp 55–58.
Han MS (1962): Unvoicing of vowels in Japanese. Nihon Onsei Gakkai 10:81–100.
Harrington J (2012): The relationship between synchronic variation and diachronic change; in Cohn A, Fougeron C,
Huffman M (eds): Handbook of Laboratory Phonology. Oxford, Oxford University Press, pp 321–332.
Harrington J, Fletcher J, Roberts C (1995): An analysis of truncation and linear rescaling in the production of
accented and unaccented vowels. J Phonet 23:305–322.
Iskarous K, McDonough J, Whalen DH (2012): A gestural account of the velar fricative in Navajo. Lab Phonol
3:195–210.
Jannedy S (1995): Gestural phasing as an explanation for vowel devoicing in Turkish. Ohio State Univ Work Papers
Ling 45:56–84.
Kent R, Kim Y (2003): Toward an acoustic typology of motor speech disorders. Clin Ling Phonet 17:427–445.
Kondo M (1997): Syllable constraints on Japanese vowel duration and intensity; in Groupe Acoustique,
Apprentissage et Interprétation (eds): La voyelle dans tous ses états. Nantes, Université de Nantes, pp
32–37.
Kondo M (2005): Syllable structure and its acoustic effects on vowels in devoicing environments; in van de Weijer
J, Nanjo K, Nishihara T (eds): Voicing in Japanese. Berlin, Mouton de Gruyter, pp 229–245.
Labov W (1972): Sociolinguistic Patterns. Philadelphia, University of Pennsylvania Press.
Lemle M (1966): Phonemic System of the Portuguese of Rio de Janeiro; thesis, University of Pennsylvania.
Lindblom B (1963): Spectrographic study of vowel reduction. J Acoust Soc Am 35:1773–1781.
Meneses F (2012): As Vogais Desvozeadas No Português Brasileiro: Investigação Acústico-Articulatória; master
degree thesis, Unicamp, Campinas.
Meneses FO, Albano E (2013): From devoicing to apocope: an acoustic study of poststressed high vowel lenition in
Brazilian Portuguese. Phonetics and Phonology in Iberia, Universidade de Lisboa.
Ohala JJ (1981): The listener as a source of sound change; in Masek CS, Hendrick RA, Miller M (eds): Papers from
the Parasession on Language and Behavior. Chicago, Chicago Linguistics Society, pp 178–203.
Ohala JJ (1993): The phonetics of sound change; in Jones C (ed): Historical Linguistics: Problems and Perspectives.
London, Longman, pp 237–278.
Pagel D (1993): Contribuição para o estudo das vogais orais inacentuadas no português brasileiro. Cad Estud Ling
25:85–99.
Ramus F, Dupoux E, Mehler J (2003): The psychological reality of rhythm classes: perceptual studies. 15th ICPhS,
Barcelona, pp 337–342.
Rolo M, Mota J (2012): Um estudo sociolinguístico sobre o apagamento de vogais finais em uma localidade rural da
Bahia. Londrina, SIGNUM: Estudos Linguísticos, No 15/1, pp 311–334.
Romero J (1996): Articulatory blending of lingual gestures. J Phonet 24:99–111.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 3/11/2016 7:14:25 PM

136 Phonetica 2015;72:121–137 Meneses/Albano

DOI: 10.1159/000439599
Downloaded by:
Saltzman E, Munhall K (1989): A dynamical approach to gestural patterning in speech production. Ecol Psychol
1:333–382.
Souza RFN (2012): A Redução de Vogais Altas Pretônicas no Português de Belo Horizonte: Uma Abordagem
Baseada na Gradiência; master degree thesis, UFMG.
Smith CL (2003): Vowel devoicing in contemporary French. J French Lang Stud 13:177–194.
Taylor M (1994): The Interaction of Vowel Deletion and Syllable Structure Constraints; PhD thesis, Simon Fraser
University, Burnaby.
Tsuchida A (1997): Phonetics and Phonology of Japanese Vowel Devoicing; PhD thesis, Cornell University, Ithaca.
Vaissière J (2009): Articulatory modeling and the definition of acoustic-perceptual targets for reference vowels.
Chin Phonet J 2:22–33.
Van Bergem D (1995): Acoustic and Lexical Vowel Reduction, in Studies in Language and Language Use 16; PhD
thesis, University of Amsterdam.
Viaro ME (2005): Algumas considerações acerca do português falado quatrocentista e quinhentista. Brasília, Papia,
vol 15, pp 80–101.
Viegas MC, Oliveira AJ (2008): Apagamento da vogal átona final em Itaúna/MG e atuação lexical. Rev Abralin
7:119–138.
Wheeler MW (2007): La síncope i l’apòcope en la història del català: una aproximació des de l’Optimitat; in
Wheeler MW (ed): Morfologia i fonologia catalana i romànica: estudis diacrònics. Valencia, Universitat de
València.
Zsiga EC (2000): Phonetic alignment constraints: consonant overlap and palatalization in English and Russian. J
Phonet 28:69–102.

Siriraj Medical Library, Mahidol University

198.143.39.97 - 3/11/2016 7:14:25 PM

Final Poststressed Vowel Devoicing in Brazilian Phonetica 2015;72:121–137 137

Portuguese DOI: 10.1159/000439599
Downloaded by: