Content uploaded by Christoph Gabriel
Author content
All content in this area was uploaded by Christoph Gabriel on May 02, 2020
Content may be subject to copyright.
57
Mario Ruiz Moreno & Christoph Gabriel
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
Abstract. Pomerano is a scarcely researched variety of Low German, which was
brought to Southern Brazil by large groups of immigrants from former Pomerania in the
1850s. Today, the language is still used in situations of informal communication in the
respective communities, along with the majority language, Brazilian Portuguese (BP).
The long-lasting contact between Pomerano and BP has left several traces in both lan-
guages at all linguistic levels, including segmental phonology. Our study examines the
Voice Onset Time (VOT) patterns of stop consonants in the two languages spoken by
bilinguals. Control data gathered from monolingually raised speakers of BP are taken
into account. The language pair under investigation displays the typical Germanic vs.
Romance contrast, in that the phonological fortis-lenis opposition is phonetically real-
ized by means of a long vs. short lag distinction in Pomerano, whereas BP exhibits a
voicing lead for the lenis stops and a short lag for their fortis counterparts. Based on
production data collected using a picture naming task in Pomerode (Santa Catarina,
Brazil), it is shown that both elderly and younger speakers show the expected Germanic
contrast in Pomerano. Regarding BP, the monolingual speakers present the typical Ro-
mance contrast, while the younger bilinguals display a mixed system in that they pre-
voice /b d ɡ/, but aspirate the voiceless stops. The older bilinguals, finally, use the same
Germanic contrast in both of their languages, thus showing massive transfer from Pom-
erano to BP. Referring to Labov’s concept of overt vs. covert prestige, we argue that the
latter, as becomes manifest in the stressing of a ‘Germanic’ feature of pronunciation,
operates within both generations, although its effect can be felt with less intensity
among the younger bilinguals.
1. Introduction
Research on diaspora varieties of German has largely concentrated on phe-
nomena other than pronunciation (cf. the volume edited by Földes 2019 for
an overview). This also holds true for varieties of Central and Low German
such as Hunsrückisch or Pommerisch (henceforth: Pomerano), which were
brought to Southern Brazil by immigrants from former Pomerania in the
early 19th century and which nowadays are still spoken in the respective bi-
lingual communities along with Brazilian Portuguese (BP). Only few em-
pirical studies have addressed specific features of Pomerano segmental
phonology, among them the production of stop consonants (cf. Schaeffer/
Gabriel, Christoph / Pešková, Andrea / Selig, Meisenburg (eds.): Contact, variation and change. Studies in
honor of Trudel Meisenburg (Studienreihe Romania 35). Berlin: Erich Schmidt, 57–77.
Ruiz Moreno & Gabriel
58
Meireles 2011; Postma 2019: 32–34). These segments constitute a fruitful
testing ground for the investigation of cross-linguistic influence (CLI) be-
tween the community language, Pomerano (Germanic), and the surround-
ing language, BP (Romance), since the characteristics of the lenis and fortis
plosives, i.e. /b d ɡ/ and /p t k/, respectively, fundamentally differ from one
another in the two languages (cf. 2.2).1 The present contribution aims to fill
a research gap in this respect by not only analyzing the acoustic features of
these segments in both of the languages spoken by the bilinguals as well as
in monolingual BP control data recorded by speakers from the same region,
but also by taking into account different age groups (cf. Section 3). The in-
terpretation of the results within the context of the Labovian distinction of
so-called overt vs. covert prestige (cf. Labov 1966/2006) will hopefully add
to our understanding of the interdependencies between speakers’ attitudes
and beliefs and sound production in multilingual communities.
Our paper is organized as follows. In a first step, the reader is provided
with the relevant background information regarding both the sociohistorical
context of Pomerano in Southern Brazil and the specific feature of stop
production we analyze, i.e. the VOT patterns of Romance and Germanic
languages (cf. Section 2). The following section is devoted to the presenta-
tion of the empirical study and the respective results (cf. Section 3). Section
4, finally, offers an outlook and some concluding remarks.
2. Background
This section offers an overview of the research background. First, the read-
er is provided with the relevant information on the Low German diaspora
variety under discussion, Pomerano, focusing on both its history and the
current contact situation with BP in Southern Brazil (cf. 2.1). Section 2.2,
then, concentrates on the production of stop consonants in the two language
families the varieties addressed belong to, i.e. Germanic and Romance. In a
last step, we offer a brief survey of the recent literature on VOT patterns in
the speech of multilinguals (cf. 2.3).
1 CLI also shows up in other areas of Pomerano phonology. For instance, front
rounded vowels such as [y] are occasionally replaced by their unrounded
counterparts in young BP-dominant speakers’ productions, e.g. [myts] ~
[mits] ‘cap’. A further example is the spreading of the nasal feature to adja-
cent vowels following the BP model as in e.g. [ʃlãŋ] ‘snake’ (cf. Schaeffer/
Meireles 2014: 52; Postma 2019: 68–69).
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
59
2.1 Pomerano: a Low German variety spoken in Southern Brazil
Pomerano is a Low German variety originally spoken in the extinct prov-
inces of Eastern Hither Pomerania (German: Ostvorpommern) and Farther
Pomerania (German: Hinterpommern; cf. Rost 2008; da Silveira 2010;
Postma 2019: 1–9). These territories nowadays belong to Poland and virtu-
ally host no German-speaking populations. As early as in 1824, the first
Pomerano speakers came to Brazil; however, the largest contingents ar-
rived during the period of 1850–1870 and settled in various locations
throughout the country (cf. Beilke 2013). Tressmann (2007; 2011) indicates
that Pomerano is nowadays spoken in the Brazilian states of Espírito Santo,
Minas Gerais, Rondônia, Santa Catarina and Rio Grande do Sul, as a result
of the immigration of around 30,000 speakers (cf. Struck 1992: 56). Tress-
mann (2011) reports a number of about 300,000 descendants of Pomerani-
an immigrants in Brazil, though without differentiating between those who
actually speak Pomerano and those who have abandoned the Germanic va-
riety spoken by their ancestors.
Immediately after their arrival in Brazil, Pomerano speakers got in
close contact with speakers of other Low German varieties as well as with
speakers of German origin outside the Low German-speaking domain, giv-
en that German settlers of different origins got mixed during the foundation
of many cities (cf. Willems 1946: 61–78 for a detailed description). As de-
scribed in Tressmann (1998), it was not infrequent that some contingents of
Pomerano speakers were outnumbered by those of Hunsrückisch (a form of
West Central German spoken in, for instance, Rio Grande do Sul; cf. Al-
tenhofen 1996) or vice versa, which resulted in a linguistic shift toward the
prevailing linguistic form in the area. Additionally, these speakers were to a
greater or lesser extent as well in contact with Standard German (so-called
Hochdeutsch), which was used as the language of instruction in locally
supported schools (cf. Fossile 2010) and in the local press.2 Moreover,
German has always been highly valued given its status as a church lan-
guage within this protestant community.
The situation described so far was quite stable for several decades. As
noted in Vandresen/Rodrigues (2008), there were numerous monolingual
communities until 1937, since mixed marriages with the local catholic pop-
ulation were difficult for religious reasons. This might explain the fact that
as late as in 1940, 97% of the population of the city of Blumenau (located
at a distance of 31 kilometers from Pomerode) declared to use “German” as
their home language (cf. Fritzen 2008). However, Portuguese gradually
2 A standardized orthography for Pomerano was established only recently by
Tressmann (2006); cf. Postma (2019: 31) for an overview.
Ruiz Moreno & Gabriel
60
gained new domains, since more and more immigrants from other parts of
Brazil settled in the Southern regions, thereby consistently altering the so-
ciolinguistic dynamics of the area.
During Getúlio Vargas’ dictatorship (1937–1945), the Pomerian lan-
guage was severely persecuted (cf. da Silveira 2010). This period was char-
acterized by the government’s attempts to achieve national homogenization
and a complete lack of respect for other languages, which lead many par-
ents to use solely Portuguese with their children (cf. Maltzahn 2018). It
comes as no surprise that with the events in World War II the sociolinguis-
tic status of Pomerano and (Standard) German, which were both perceived
as the language of the enemy, was weakened even more in Brazil, ending in
a complete ban of both German language classes from schools and the use
of German in educational institutions. Under these circumstances, preserv-
ing any language other than Portuguese was already a hard-won success,
but Pomerano-(Standard) German-Portuguese trilingualism was particular-
ly challenging to maintain. The current general situation of the language
differs vastly from one place to another. In some communities, German is
on decline, whereas Pomerano is still much spoken (cf. Vandresen/Rodri-
gues 2008); conversely, in Pomerode, where our data collection took place,
the situation has been reversed and Pomerano has lost its predominance
over German during the last decades (cf. Maltzahn 2018).
2.2 Stop consonants in Germanic and Romance
Stop production varies considerably between languages and language fami-
lies. A salient feature that differentiates the language-specific realizations
of plosives is the so-called Voice Onset Time (VOT), which refers to the
time elapsing between the release of the stop and the voicing onset of the
following vowel (cf. Lisker/Abramson 1964: 389; Cho/Ladefoged 1999;
Gabriel et al. 2013: 60–62). Three types of VOT are usually distinguished:
(1) pre-voicing or voicing lead, i.e. the vibration of the vocal folds starts prior
to the release, which yields negative VOT values,
(2) short lag, i.e. voicing starts with or shortly after the release, which yields a
short positive VOT and, finally,
(3) long lag, i.e. voicing starts considerably after the release, yielding a long
positive VOT.
Most languages present a binary phonological contrast (lenis-fortis) which
is cross-linguistically associated with different VOT types. Tab. 1 summa-
rizes the characteristics of stop consonants and their respective VOT pat-
terns in Germanic and Romance.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
61
pre-voicing
(negative VOT)
short lag
(short positive VOT)
long lag
(long positive VOT)
Germanic [b̥ d̥ ɡ̊] <b d g>
lenis (devoiced)
[
p
hthkh] <p t k>
fortis (voiceless, aspirated)
Romance [b d ɡ] <b d g>
lenis (fully voiced)
[
p
t k] <p t k>
fortis (voiceless)
Tab. 1. VOT patterns and orthographic representations of stop consonants in Germanic
(Pomerano, German) and Romance (BP).
Within the Germanic family, which includes both Standard German and
Pomerano, most languages phonetically express the lenis-fortis opposition
by short-lagged unaspirated (devoiced) [b̥ d̥ ɡ̊] on the one hand and long-
lagged [ph th kh] on the other. 3 In German, aspiration does not occur in con-
sonant clusters and is strongest before stressed vowels (cf. Wiese 1996:
270). Recent studies reported VOTs of 76 ms (cf. van de Weijer/Kupisch
2015: 417) and 78.5 ms (cf. Lein et al. 2016: 739) as average values for
/p t k/. For /k/, where VOT is highest, former studies reported dramatically
lower VOT durations of 37–67 ms (cf. Stock 1971; Fischer-Jørgensen
1979), but these values might have been influenced by both diatopic varia-
tion and the materials analyzed, e.g. the fact that no distinction was made
between onsets of unstressed and stressed syllables (cf. Lein et al. 2016:
735 for an overview). As compared to German, empirical research on stop
production in Pomerano, which also exhibits the typical Germanic contrast,
is quite scarce. Postma (2019: 32) points out that “aspiration is […] weaker
than in Standard German”, though without presenting any VOT measure-
ments. Schaeffer/Meireles (2011: 4013) report for /p t k/ long lag VOT
values of 52, 54 and 72 ms, as expected;4 for the lenis stops, their meas-
urements present some inconsistencies in that /b/ is pre-voiced (-26 ms),
while /d ɡ/ present short lag VOTs of 17 and 3 ms, respectively.
The literature on Brazilian Portuguese (BP), finally, reports the ex-
pected contrast of truly or pre-voiced lenis and short-lagged fortis stops (cf.
Melo et al. 2014: 491, using lab data, both word-initially and -medially).
3 Not all Germanic languages follow this pattern: Swedish has a mixed sys-
tem, which comprises pre-voiced lenis and aspirated fortis plosives (cf. Hel-
gason/Ringen 2008). Dutch, finally, patterns with the Romances family in
exhibiting truly voiced /b d ɡ/ and unaspirated /p t k/ (cf. van Alphen/Smits
2004). Dialectal variation in German is addressed later in this section.
4 The values for Pomerano voiceless stops reported by Bandeira/Zimmer
(2011: 93) slightly differ from the ones given by Schaeffer/Meireles (2011),
i.e. 51 ms for /p/, 42 ms for /t/ and 76 ms for /k/. The fact that the value for
the alveolar stop is (atypically) lower than the one for the bilabial plosive is
not commented on by the authors.
Ruiz Moreno & Gabriel
62
Reis/Nobre-Oliveira’s (2007: 402) study on the learning of L2 English by
BP natives is confined to voiceless stops and reports for the learners’ L1
the expected short-lagged productions (i.e. 17.27 ms for /p/, 23.55 ms for /t/
and 46.55 ms for /k/).
As shown in Tab. 1, the Germanic realization of the lenis plosives
roughly patterns with the production of their fortis counterparts in Ro-
mance. Transfer of VOT patterns as frequently occurs in situations of lan-
guage contact and foreign language learning can thus provoke misunder-
standings among speakers. For instance, an intended production of the
Portuguese verb form bato /ˈbato/ ‘I hit’ (1SG.PRS of bater), pronounced as
[ˈb̥atu]/[ˈpatu] by a learner whose L1 is a Germanic variety, might be misin-
terpreted by Portuguese listeners as pato ‘duck’. Note that misinterpreta-
tions of this kind may also occur between speakers of varieties of the same
language which differ from one another regarding their VOT patterns, as
e.g. Northern Standard and Vienna German.5 However, since isolated pro-
ductions of single words rarely occur in natural conversation and the inter-
pretation of an utterance is highly supported by the respective (linguistic
and extra-linguistic) context, communication is usually not disturbed by
such instances of CLI. However, stop production crucially contributes to
the perception of both foreign accent (in non-native speech) and the per-
ceived ‘otherness’ of varieties other than the one which is spoken by the
hearer. This aspect will be discussed in more detail in 3.4.
2.3 VOT patterns in multilingual speakers
At least since the publication of Laeufer’s (1997) seminal paper, the char-
acteristics of VOT patterns in the production of speakers who use more
than one language regularly cannot be claimed to be an understudied field
(cf. Kupisch/Lleó 2017 for a summary of recent research). While simulta-
neously bilingual children who acquire two languages that differ from one
another concerning their VOT patterns (e.g. Spanish and English) tend to
separate the phonologies of their languages with respect to this feature (cf.
Deuchar/Clark 1996), evidence for transfer of VOT patterns was found in
many other multilingual constellations. For instance, Fowler et al. (2008)
investigated the production of voiceless stops in sequential bilinguals of
Canadian French and English, showing that the VOT values of the lan-
5 Cf. e.g. Gepäck ‘luggage’, which is pronounced in Vienna German with a
shorter VOT for the word-medial fortis plosive /p/ than in Northern German
(cf. Moosmüller et al. 2015: 341), and might thus be misinterpreted as an in-
stance of intended Gebäck ‘biscuits, pastries’ by Northern German listeners.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
63
guage acquired later were considerably influenced by the speakers’ L1. CLI
was also reported to occur in speakers of a heritage language (HL). As
shown by Kupisch/Lleó (2017), both heritage speakers of German living in
Italy and speakers of Italian as a HL who live in Germany had developed
intermediate VOT values for /k/ in their respective HL, situated between
the ones reported for German/Italian spoken by monolinguals. Interesting-
ly, the authors also showed that, at least in some of their participants, not
only the HL, but also the surrounding language was affected by VOT trans-
fer. Similarly, Post/Jones (this volume) showed that English-French late bi-
linguals living in the UK for at least 10 years have developed compromise
values for /t/, situated between the durations of the two monolingual varie-
ties, English and French. Finally, CLI in the realm of stop production is
widely attested in foreign language (FL) learning, e.g. in English learners
of Spanish, who transfer the long-lagged realizations of voiceless plosives
from their L1 to the FL (cf. e.g. Zampini 2014) and, though to a lesser ex-
tent, also in heritage speakers of Spanish residing in English- or German-
speaking environments (cf. Amengual 2012; Ruiz Moreno 2019). Positive
transfer of VOT values from the HLs Russian and Turkish, respectively, in
adolescent multilingual learners who acquire French as a FL in the German
educational context was reported by Gabriel et al. (2018).
3. Empirical study
In the following, we first outline our research questions (3.1) and the meth-
odology of our study (3.2). In a next step, we present the results (3.3) and
discuss them in the context of Labovian sociolinguistics (3.4).
3.1 Research questions
Our research had a double goal. First, it was aimed to describe the VOT
patterns of the two languages spoken by a group of Pomerano-BP bilingual
speakers from Pomerode (Santa Catarina) and to thereby answer the ques-
tion of whether phonic transfer occurs within this community. Second, giv-
en the growing influence of BP among younger speakers and the circum-
stances of linguistic repression typical of the mature speakers’ childhood,
we aimed to determine whether two different generations diverted in their
VOT patterns. This aspect had not been investigated before, since the
scarce previous studies on Pomerano VOT focused only on younger speak-
ers (aged 15–24 in Schaeffer/Meireles 2011 and 8–10 in Bandeira/Zimmer
Ruiz Moreno & Gabriel
64
2011), which contrasts with the wider age range of the speakers recorded in
the context of our study.
3.2 Methods, speakers and materials
A total of 18 participants took part in this research: 4 monolingual speakers
of the local BP variety (mean age: 39; each 2 males and females; group M),
7 bilinguals younger than 70 (mean age: 57; 6 males, 1 female; group B≤70)
and, finally, 7 bilinguals older than 70 (mean age: 81; 4 males, 3 females;
group B>70).6 All participants were contacted with the help and mediation
of a local, who also introduced the instructor to the participants and was
present at each interview, which created a friendly atmosphere. Since there
is a wide consensus in the literature on allochthonous languages in Brazil
that the dictatorship had a deep impact on these languages, we decided to
split the participants in these two groups instead of running correlations be-
tween their VOT values and the ages of our participants to test whether the
deep impact that the dictatorship had on aspects such as language use and
language attitudes was also reflected in their VOT production. The cutting-
point was established at the age of 70 because the data were collected in
2015 and the dictatorship ended in 1945; consequently, the participants
born after the dictatorship fell into one group and those born before or dur-
ing it into another one. All of the participants had the option to receive the
instructions in German or Portuguese and they typically chose German.
The data collection was carried out in Pomerode (Santa Catarina) in Jan-
uary 2015.
Before proceeding to the recordings, the instructor and the participants
chatted informally for a while, and then the instructions were given. The
data were collected using a picture naming task. Participants were ex-
plained that they would be shown a series of pictures on a laptop screen
6 Regarding possible gender differences in VOT production, the picture that
emerges from the literature is inconsistent. Some studies report longer VOTs
in male speakers, e.g. for Florentine Italian, where respective gender differ-
ences seem to be linked to register and socio-economic status of the speakers
(cf. Piccardi 2017); others, however, found greater VOT values in female
speech (cf. Swartz 1992 for US English). Still others found sex differences in
VOT production for young children, but not in adult speech (cf. Karlsson et
al. 2004 for Swedish). To our knowledge, no empirical study reports system-
atic sex or gender-based differences in VOT measurements for BP. We thus
conclude that the male-biased gender distribution of our younger bilingual
group (86% males) does not question its comparability with the non-biased
group of the older bilinguals presenting an equal ratio of males and females.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
65
and that they were required to utter just one word to name the depicted
item. They were also given the opportunity to choose the language in which
they wished to be recorded first (i.e. either Pomerano or BP). Finally, they
were explained that the research was by no means a kind of exam to test
their knowledge of Pomerano or BP since no right or wrong answer was
expected; they were solely required to speak as naturally as possible. 18
items were recorded in Pomerano and 19 in BP, using a Marantz profes-
sional PMD671 sound recorder. The items contained /b d ɡ/ and /p t k/ in
the onset position of word-initial stressed syllables in both languages. Bauk
‘book’, tung ‘tongue’, täne ‘teeth’, dek ‘blanket’ or pol ‘onion’ exemplify
this in Pomerano, whereas gota ‘drop’, carro ‘car’, pera ‘pear’, dado ‘dice’
or dedo ‘finger’ in BP. Both words with initial stops followed by front vs.
back vowels were selected in a balanced way.
In rare occasions, participants uttered non-targeted words. In these cas-
es, they were first asked to use a synonym. When this proved unsuccessful,
they were told the expected word and asked to repeat it. Additionally, it
must be noted that, given the advanced age of the older group of bilinguals,
most but not all items were successfully recorded by each speaker. In some
cases, for instance, participants refused to utter the Pomerode word denot-
ing a glass instead of the one meaning ‘cup’. In others, they only murmured
the target item even when asked to repeat it audibly. The VOT measure-
ments were done using Praat (Boersma/Weenink 2014); an example is giv-
en in Fig. 1.
Fig. 1. VOT measurement of the Pomerano item dek ‘blanket’ produced with a short lag
([d̥ɛk]) in Praat (VOT duration: 12.8 ms).
Ruiz Moreno & Gabriel
66
3.3 Results
This section gives an overview of the results. Fig. 2 and Tab. 2 represent
the VOT measurements for the BP data.
Fig. 2. Mean VOT values (ms) for the BP items produced by three groups of speakers,
i.e. older bilinguals (B>70), younger bilinguals (B≤70) and BP monolinguals (M).
B>70 B≤70 M
VOT SD VOT SD VOT SD
/
p
/ 38 14.00 27 6.32 15 1.47
/t/ 37 5.40 31 1.60 21 1.97
/k/ 51 11.07 49 9.62 46 8.29
/b/ -4 30.45 -76 24.76 -66 9.89
/d/ 5 24.79 -83 27.23 -60 11.74
/ɡ/ 16 15.47 -63 24.58 -58 20.60
Tab. 2. Mean VOT values (ms) and standard deviations (SD) for the BP items produced
by three groups of speakers, older bilinguals (B>70), younger bilinguals (B≤70) and BP
monolinguals (M); all differences statistically significant (Kruskal-Wallis, Mann-Whitney)
between groups except for /k/.
As can be seen in Fig. 2 and Tab. 2, the older bilinguals (B>70) clearly fol-
low a different path than the younger bilinguals (B≤70) and the BP mono-
linguals (M) in that they produce Portuguese lenis stops with short lag, i.e.
as they would in Pomerano (cf. below). Fully voiced instances of /b d ɡ/
occur only rarely in this group; two of the B>70 speakers produced a single
voiced token for /b/. These two outliers were nevertheless included in the
mean, since they correspond to the target form in BP. The most common
phonetic realization for /b/, however, in group B>70 is a (Germanic) short-
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
67
lagged voiceless (or: devoiced) stop, but given the restricted data bases,
two tokens with negative VOTs are sufficient to distort the mean. Lastly,
the fortis tokens in this group are also clearly more aspirated than those
produced by the monolinguals (M) and the younger bilinguals (B≤70), ex-
cept for /k/.
In turn, the younger generation of bilingual speakers (B≤70) represents
an intermediate position. They pattern with the monolinguals in clearly pre-
voicing the lenis plosives /b d ɡ/, but they differ from them in their VOTs
for /p t k/, which are considerably more aspirated, although to a lesser ex-
tent than those produced by the older bilinguals (B>70).
Statistical analyses were done in SPSS. A battery of Mann-Whitney
tests confirmed that the older bilinguals (group B>70) differ significantly
from the monolinguals (group M) for every single phoneme except for /k/,7
whereas the younger bilinguals (B≤70) only presented significant differ-
ences for /p/ and /t/, when compared to the BP monolinguals (group M).
Finally, when comparing both bilingual groups, significant differences
were found for /b d ɡ/, /p/8 and /t/.
The results for Pomerano are represented in Fig. 3 and Tab. 3, below.
Fig. 3. Mean VOT values (ms) for the Pomerano items produced by two groups of
speakers, i.e. older (B>70) and younger bilinguals (B≤70).
7 As can be seen in Fig. 2, above, there are indeed the same differences which
follow the pattern found for /p/ and /t/, but given the limited number of to-
kens, differences need to be considerably large to reach statistical signifi-
cance. In this very case, our BP monolinguals seem to have aspirated /k/ to a
somewhat greater extent than normally. Note that, nevertheless, their VOT
values were lower than those produced by the bilinguals.
8 The p-value for /p/ is 0.097, which we consider nonetheless as a significant
difference since we set α at 0.1 as suggested by Larson-Hall (2010: 98–99)
for statistical tests applied in the field of linguistics.
Ruiz Moreno & Gabriel
68
B>70 B≤70
VOT SD VOT SD
/
p
/ 51 13.49 43 15.49
/t/ 56 8.28 51 12.69
/k/ 71 11.83 71 11.75
/b/ 14 4.90 14 3.06
/d/ 21 7.91 25 9.95
/ɡ/ no data
Tab. 3. Mean VOT values (ms) and standard deviations (SD) for the Pomerano items
produced by two groups of speakers, i.e. older (B>70) and younger bilinguals (B≤70);
differences between groups not significant. No data were obtained for /ɡ/, which was
produced as a velar approximant by all speakers.
Not unexpectedly, no measurements of [ɡ] could be obtained, since no
speaker produced a stop for initial <g>, but rather a velar approximant or
glide as in the study by Schaeffer/Meireles (2014); cf. also Postma (2019:
32). As for the remaining stops, no statistical differences between the two
bilingual groups were found, which automatically discards the possibility
that the differences in their BP may be attributable to existing differences
in their Germanic L1. Both groups consistently produced voiceless unaspi-
rated tokens for the lenis stops, and the fortis counterparts were clearly as-
pirated. In other words, whereas [±voicing] is the contrasting feature for BP
stops, [±aspiration] is the feature whereby the phonological contrast is
maintained in Pomerano.
It is worth noting that there were four instances (two of them produced
by the same speaker) of pre-voiced lenis stops, i.e. [b d ɡ], which were not
included in the calculus of the group means to prevent extreme distortions.
The apparent asymmetry in the treatment of these outliers and those found
in the BP data is justified by the fact that, in this occasion, we are dealing
with clearly non-target forms. Furthermore, the presence of voicing in these
tokens cannot be exclusively attributed to the long-lasting contact of Pom-
erano with BP, since even among Low German speakers in Germany the
presence of voicing is not rare (cf. Ouddeken 2016); it is simply phonologi-
cally redundant.
Summing up, the phonological contrast for the BP and Pomerano stops
in each group can be described as follows.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
69
BP Pomerano
pre-voicing short lag long lag pre-voicing short lag long lag
B>70 /b̥ d̥ ɡ̊/ /
p
ʰ tʰ kʰ/ /b̥ d̥ ɡ̊/ /
p
ʰ tʰ kʰ/
B≤70 /b d ɡ/ /
p
ʰ tʰ kʰ/ /b̥ d̥ ɡ̊/ /
p
ʰ tʰ kʰ/
M /b d ɡ/ /
p
t k/
Tab. 4. Phonological representation of stops in BP and Pomerano in three groups of
speakers, older bilinguals (B>70), younger bilinguals (B≤70) and BP monolinguals (M).
3.4 Discussion
One of our main goals was to describe the typical VOT values for this
Pomerano-BP bilingual community in order to determine whether or not
CLI at the phonic level had taken place. We also aimed to answer the ques-
tion of whether this possible transfer was subject to intergenerational varia-
tion. Such a scenario was plausible given that in the Pomerano-speaking
communities, BP is gaining more and more ground, particularly among
younger speakers, due to the penetration of BP as a result of Portuguese-
speaking immigration from other regions of Brazil and a more intense ex-
posure to BP through the mass media. The answer to both questions is
clearly affirmative: phonic transfer does occur within this bilingual com-
munity, and there are noticeable differences depending on speaker age.
The first aspect worth noting is that transfer of VOT values is unidirec-
tional: the Pomerano values of these speakers remain unaltered despite the
contact with Portuguese. As opposed to this, their Portuguese VOT values
are affected by the contact with Pomerano. The effect of this contact is
more noticeable for the older generation, since they have substituted the
typical Romance contrast in terms of [±voicing] by their L1 contrast based
on [±aspiration], which is hardly surprising given how unbalanced their
linguistic input was. The younger generation, by contrast, had much contact
with BP since very early childhood, which can be noticed in their ability to
produce truly voiced lenis stops in BP; however, their BP fortis stops are
more aspirated than those produced by our monolinguals. This is interest-
ing, because their mixed VOT system in BP is perfectly functional from a
communicative point of view. Seen from this angle, it is phonologically
important, unlike the older bilinguals, they avoid devoicing in /b d ɡ/ (i.e.
[b̥ d̥ ɡ̊]), since devoiced realizations are easily perceived as /p t k/ by BP
listeners in spite of the tenseness differences. This perceptual overlap caus-
es misspellings in BP among Pomerano speakers (cf. Benincá 2008;
Schaeffer 2010). In contrast, their [pʰ tʰ kʰ] tokens sound certainly odd to
the Brazilian ear used to hearing [p t k], but are perceived as /p t k/ tokens
after all.
Ruiz Moreno & Gabriel
70
Summing up, whereas the older bilinguals (B>70) make use of a single
VOT system in both languages, leaving their Pomerano VOTs unaltered,
the younger bilinguals (B≤70) present a different VOT system in both of
their languages. Their Pomerano stops are fully canonical, whereas they
produce canonical lenis stops in BP, but aspirated fortis ones, most likely
due to partial transfer of Pomerano to BP.
It is also of interest to compare the VOT values determined in this
study with those obtained in previous research. The data for BP monolin-
guals in Schaeffer/Meireles (2011), in Reis/Nobre-Oliveira (2007) and Me-
lo et al. (2014) with speakers from Espírito Santo, Santa Catarina and Rio
Grande do Sul, respectively, are in total accord with ours. As for the VOT
values in BP of Pomerano-Portuguese bilinguals, there are some differ-
ences between our study and the other two (cf. Tab. 5). Bandeira/Zimmer
(2011) only report their /p t k/ values, which are clearly more aspirated
than those of our Santa Catarina bilinguals and those of Schaeffer/Meireles
(2011), who collected their data in Espírito Santo. The community studied
by Bandeira/Zimmer (2011) might be less exposed to BP, but it is also
plausible to assume that this phonic difference is simply due to the fact that
their participants were aged 8–10 and thus had had less time to learn to
modulate possible L1 interferences. As for /b d ɡ/, we can see that
Schaeffer/Meireles’ bilinguals, aged 15–25, pattern with our younger group
(B≤70) in that they also present a mixed model with pre-voiced /b d ɡ/ and
aspirated fortis stops. Their lenis VOT values are nonetheless closer to ze-
ro, which most likely suggest the presence of outliers who produce [b̥ d̥ ɡ̊]
tokens.
Regarding Pomerano VOT, it is worth highlighting the clear homoge-
neity among studies for the /p t k/ values. For the VOT comparison for
/b d ɡ/, the only available data are those of Schaeffer/Meireles (2011).9
Their values for /d/ are in line with ours, but the negative value for /b/ and
the positive but close to zero for /ɡ/ are rather intriguing and seem to indi-
cate the presence of outliers with voiced tokens for these two stops. As not-
ed before, this is not rare even among Low German speakers with no con-
tact with Romance languages (cf. Ouddeken 2016), but it is definitely not
the canonical form. Furthermore, it should be pointed out that their data are
based on only four participants.
9 Note that their data were elicited using the carrier-sentence Ik sai … nuu ‘I
see … now’ (cf. Schaeffer/Meireles 2011: 4012). Given the influence of the
previous vowel, voicing can occur naturally as a result of assimilation in
their tokens, but not in ours, in which the target segment appeared in abso-
lute initial position.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
71
study B/Z (2011) S/M (2011) present study
state of data collection RS ES SC
age of participants 8–10 15–25 ≤70 >70
number of participants 20 4 7 7
language BP Pom. BP Pom. BP Pom. BP Pom.
/
p
/ 50 51 35 52 27 51 38 43
/b/ – – -25 -26 -76 14 -4 14
/t/ 59 42 32 54 31 56 37 51
/d/ – – -24 17 -83 21 5 25
/k/ 67 76 49 72 49 71 51 71
/ɡ/ – – -2 3 -63 – 16 –
Tab. 5. Comparison with earlier empirical studies on VOT in Pomerano-BP bilinguals
(B/Z (= Bandeira/Zimmer) 2011; S/M (= Schaeffer/Meireles) 2011). RS = Rio Grande
do Sul, ES = Espírito Santo, SC = Santa Catalina.
Finally, a tentative explanation for the intergenerational variation found in
our data will be provided. However, given the modest number of partici-
pants and tokens analyzed, any interpretation must be taken with due cau-
tion. There are – at least – two logical explanations for the age differences
in BP VOT values among our bilinguals, which nonetheless are not mutual-
ly exclusive. One is that they received different amounts of linguistic input
in BP; the other one rather refers to the speakers’ self-concept, i.e. to dif-
ferent sociolinguistic values, which are mirrored in the phonetic forms they
opt for.
Although no self-reports on current or past use of BP were collected for
this research, we can be sure, in light of the linguistic and demographic
changes in Pomerode and Brazil, that, generally, the older generation has
been less exposed to BP than the younger one. However, it must be empha-
sized that the input they received was more than enough to fully acquire
this language. Additionally, bearing in mind the dates of birth of the older
bilinguals (1927, 1929, 1931, 1933, 1936, 1937 and 1943), and even admit-
ting that some of them may have received limited schooling, it seems plau-
sible to assume that most of them received considerable BP input before
puberty (except, perhaps, the two oldest speakers), since German was
banned from schools during Getúlio Vargas’ dictatorship (1937–1945; cf.
Section 2, above). Even admitting that BP has not been their dominant lan-
guage, there is no doubt that most of them must have received considerable
input at a very early age, which, in principle, should be enough so that most
of these older speakers acquired the BP VOT patterns, at least irregularly,
i.e. presenting canonical as well as non-canonical realizations. However,
the production of their BP stops was astonishingly constant in that they
Ruiz Moreno & Gabriel
72
produced no voiced /b d ɡ/ and always aspirated their /p t k/ tokens, which
leaves the door open to alternative or complementary explanations.
Such an explanation, as advanced before, must probably be linked with
differences in the sociolinguistic values of two age groups. In any linguistic
community, there is always a constant tension between so-called covert and
overt prestige. As Labov showed in his seminal sociolinguistic work on
The social stratification of English in New York (1966/2006), many speak-
ers are fully aware of the linguistic forms that convey more prestige in their
societies and, yet, employ other ones. There is no doubt that the bilingual
speakers are fully aware of the stigmatization of the VOT transfer from
Pomerano to BP because of the mockery they receive due to their ‘German
accent’, and considering both their early exposure to BP and the total ab-
sence of voiced stops in absolute initial position as well as the regular pres-
ence of aspirated stops, it is plausible to argue that these speakers favor the
production of BP stops with a ‘German-like’ pronunciation.
They may have different reasons to do that. On the one hand, speaking
German and/or Low German is of great importance for the cultural and
ethnical identity of this group (cf. Maltzahn 2018). Thus, leaving some
traces of German influence in their BP speech could foster the cohesion of
the group or serve to express part of their identity. A complete absence of
German influence might even be regarded negatively among these speak-
ers, since they could be blamed of sounding too stilted or pedantic. On the
other hand, it is well-known in the field of sociolinguistics that when a
group is “under attack from outside” (Trudgill 2000: 13), as Low German
speakers definitely were during the repression of Getúlio Vargas (cf. Silvei-
ra 2010), the “signals of difference may become more important” (Trudgill
2000: 13) for their speakers, and producing BP stops in a (Low) German
fashion is undoubtedly a distinctive sign.
In our data, 6 participants, all of them older than 70 years, completely
transferred their Pomerano VOT pattern to BP. 5 of them lived the total of
the 8 years of Vargas’ dictatorship. The youngest member of the group
(born in 1943) only was alive during two years of this era. Interestingly,
only one of the speakers brought up during the dictatorship showed no
phonic transfer. It thus seems to be plausible that the presumably traumatic
experience of the years of repression may have triggered an unconscious
desire to preserve linguistic signs of othering. A further aspect that seem-
ingly reinforces the idea that a great deal of covert prestige is attached to
sounding German when speaking BP is the fact that the younger bilinguals
have not abandoned aspirated stops. The pressure of schooling with its em-
phasis on writing and the fact that [b̥ d̥ ɡ̊] perceptually overlaps with [p t k]
has probably led them to abandon these forms and adopt [b d ɡ] instead,
but it can hardly be assumed that these bilinguals do not perceive the dif-
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
73
ference in aspiration for their /p t k/ tokens and those of the monolinguals,
given that aspiration has been identified as a perceptually salient feature
(cf. e.g. Gordon 2018: 62–63; Chang 2018) and their mastery of Pomerano
ensures that they are sensitive to changes in aspiration. Thus, the presence
of aspirated stops in these younger speakers even in adulthood can also be
interpreted as a reflex of the covert prestige that operates within this bilin-
gual community favoring German-accented BP pronunciations that clearly
go against the overt prestige attached to canonically produced BP stops.
Lastly, it is worth remembering that Schaeffer/Meireles’ young bilin-
guals (2011) also present a combination of voiced and aspirated stops in
BP, despite being considerably younger than our younger bilinguals. They
have most likely received even more BP input than the members of our
group B≤70, which corroborates the idea that the effects of covert prestige
can also be felt in their community and that input alone can hardly explain
the differences found between groups B≤70 and B>70.
4. Concluding remarks
Our Pomerano-BP bilingual speakers present two different kinds of phonic
transfer: a first one which consists in complete transfer of VOT patterns
from Pomerano to BP (older bilinguals; group B>70) and a second one
which yields a partially fused system comprising non-canonically aspirated
fortis and canonically pre-voiced lenis plosives in BP and a canonical
(Germanic) pattern with non-aspirated and aspirated stops in Pomerano
(younger bilinguals; group B≤70). Without denying the importance of pos-
sible differences in BP input to account for the intergenerational differ-
ences, we have hypothesized that the linguistic behavior of these speakers
might be at least partially triggered by sociolinguistic motivations. As a
consequence of both the attacks that this linguistic group suffered during
Vargas’ dictatorship and their relative isolation from the BP monolinguals,
they seemingly bestow full importance upon covert prestige, i.e. they value
German-accented pronunciations positively, whilst the younger generation
occupies an intermediate step between the monolinguals and the older gen-
eration, adopting those sounds that are more important from a communica-
tive point of view, i.e. [b d ɡ], but conserving the stigmatized forms
[pʰ tʰ kʰ] that serve them as linguistic identity markers.
By no means can it be assumed that bilingual communities of alloch-
thonous languages in Brazil present no social variation, since no speaker
community is fully homogeneous. However, the sociolinguistic component
is often ignored in the studies with speakers of these communities. Hope-
fully, our study has underlined that further research on allochthonous lan-
Ruiz Moreno & Gabriel
74
guages in Brazil needs not be just framed within bilingualism studies, but
also include a sociolinguistic perspective that will enable a better under-
standing of the linguistic dynamics operating in these communities.
References
van Alphen, P. M. / Smits, R. 2004. Acoustical and perceptual analysis of the
voicing distinction in Dutch initial plosives. The role of prevoicing. Journal
of Phonetics 32, 455–491.
Altenhofen, C. V. 1996. Hunsrückisch in Rio Grande do Sul. Ein Beitrag zur Be-
schreibung einer deutschbrasilianischen Dialektvarietät im Kontakt mit dem
Portugiesischen. Stuttgart: Steiner.
Amengual, M. 2012. Interlingual influence in bilingual speech. Cognate status
effect in a continuum of bilingualism. Bilingualism. Language and Cogni-
tion 15, 517–530.
Bandeira, M. H. T. / Zimmer, M. C. 2011. A transferência dos padrões de VOT de
plosivas surdas no multilinguismo. Letras de Hoje 46, 87–95. http://revistasele
tronicas.pucrs.br/ojs/index.php/fale/article/viewFile/7503/6577 (2019-12-06).
Beilke, N. S. V. 2013. Pomerano. Uma variedade germânica em Minas Gerais.
Anais do Simpósio Internacional de Letras e Linguística 3, 1–14.
http://www.ileel.ufu.br/anaisdosilel/wp-content/uploads/2014/04/silel2013_
2440.pdf (2019-12-06).
Benincá, L. R. 2008. Dificuldade no domínio de fonemas do português por cri-
anças bilíngues de português e pomerano. PhD dissertation. Vitória: Univer-
sidade Federal do Espírito Santo. http://repositorio.ufes.br/handle/10/3689
(2019-12-06).
Boersma, P. / Weenink, D. 2014. Praat. Doing phonetics by computer. Computer
software. Version 6.0.43. http://www.praat.org.
Chang, C. B. 2018. Perceptual attention as the locus of transfer to nonnative
speech perception. Journal of Phonetics 68, 85–102.
Cho, T. / Ladefoged, P. 1999. Variation and universals in VOT. Evidence from
18 languages. Journal of Phonetics 27, 207–229.
Deuchar, M. / Clark, A. 1996. Early bilingual acquisition of the voicing contrast
in English and Spanish. Journal of Phonetics 24, 351–365.
Fischer-Jørgensen, E. 1979. Zu den deutschen Verschlußlauten und Affrikaten.
In Ezawa, K. / Rensch, K. H. Eds. Sprache und Sprechen. Tübingen: Nie-
meyer, 79–100.
Földes, C. Ed. 2019. Kontaktvarietäten des Deutschen im Ausland. Tübingen:
Narr.
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
75
Fossile, D. K. 2010. O ensino da língua alemã no sul do Brasil. Profiscientia. Pe-
riódico Multidisciplinar do IFMT (Instituto Federal de Mato Grosso) 5, 43–
60. http://www.profiscientia.ifmt.edu.br/profiscientia/index.php/profiscientia
/article/view/57/60 (2019-12-11).
Fowler, C. A. / Sramko, V. / Ostry, D. J. / Rowland, S. A. / Hallé, P. 2008. Cross
language phonetic influences on the speech of French-English bilinguals.
Journal of Phonetics 36, 649–663.
Fritzen, M. P. 2008. Línguas em contato/conflito. Em foco uma escola rural em
zona de imigração alemã no sul do Brasil. Anais do CELSUL.
http://livrozilla.com/doc/741322/línguas-em-contato-conflito (2019-12-06).
Gabriel, C. / Meisenburg, T. / Selig, M. 2013. Spanisch. Phonetik und Phonolo-
gie. Eine Einführung. Tübingen: Narr.
Gabriel, C. / Krause, M. / Dittmers, T. 2018. VOT production in multilingual
learners of French as a foreign language. Cross-linguistic influence from the
heritage languages Russian and Turkish. Revue française de linguistique ap-
pliquée 23, 59–72.
Gordon, M. 2018. Phonology. Organization of speech sounds. In Genetti, C. Eds.
2018. How languages work. An introduction to language and linguistics.
Cambridge: Cambridge University Press, 55–78.
Helgason, P. / Ringen, C. 2008. Voicing and aspiration in Swedish stops. Journal
of Phonetics 36, 607–627.
Karlsson, F. / Zetterholm, E. / Sullivan, K. P. H. 2004. Development of a gender
difference in Voice Onset Time. In Cassidy, S. / Cox, F. / Mannell, R. / Pale-
thorpe, S. Eds. Proceedings of the 10th Australian International Conference
on Speech Science & Technology. Sydney: Macquarie University, 316–321.
Kupisch, T. / Lleó, C. 2017. Voice Onset Time in German-Italian simultaneous
bilinguals. Evidence on cross-language influence and markedness. In Yavaş,
M. / Kehoe, M. / Cardoso, W. Eds. Romance-Germanic bilingual phonology.
Sheffield: Equinox, 79–98.
Labov, W. 1966/2006. The social stratification of English in New York City (2nd
ed.). Cambridge: Cambridge University Press.
Laeufer, C. 1997. Towards a typology of bilingual phonological systems. In
James, A. R. / Leather, J. Eds. Second-language speech. Structure and pro-
cess. Berlin: De Gruyter, 325–342.
Larson-Hall, J. 2010. A guide to doing statistics in second language research us-
ing SPSS. New York: Routledge.
Lein, T. / Kupisch, T. / van de Weijer, J. 2016. Voice Onset Time and global for-
eign accent in German-French simultaneous bilinguals during adulthood. In-
ternational Journal of Bilingualism 20, 732–749.
Lisker, L. / Abramson, A. S. 1964. A cross-language study of voicing in initial
stops. Acoustical measurements. Word 20, 384–422.
Ruiz Moreno & Gabriel
76
Maltzahn, P. 2018. A língua alemã como marcador de identidade étnica em Po-
merode. Pandaemonium Germanicum 21, 113–135. http://www.scielo.br/
pdf/pg/v21n33/1982-8837-pg-21-33-00113.pdf (2019-12-06).
Melo, R. M. / Mota, H. B. / Mezzomo, C. L. / Brasil, B. de C. / Lovatto, L. / Ar-
zeno, L. 2014. Acoustic characterization of the voicing of stop phones in
Brazilian Portuguese. Revista CEFAC 16, 487–499. http://www.scielo.br/
pdf/rcefac/v16n2/en_1982-0216-rcefac-16-2-0487.pdf (2019-12-06).
Moosmüller, S. / Schmid, C. / Brandstätter, J. 2015. Standard Austrian German.
Journal of the International Phonetic Association 45, 339–348.
Ouddeken, N. 2016. Voicing distinctions in the Dutch-German dialect continu-
um. Linguistics in the Netherlands 33, 106–120.
Piccardi, D. 2017. Sociophonetic factors of speakers’ sex differences in Voice
Onset Time. A Florentine case study. In Bertini, C. / Celata, C. / Lenoci, G. /
Meluzzi, C. / Ricci, I. Eds. Fattori sociali e biologici nella variazione foneti-
ca. / Social and biological factors in speech variation. Milano: Officinaven-
tuno, 83–106.
Postma, G. 2019. A contrastive grammar of Brazilian Pomeranian. Amsterdam:
Benjamins.
Reis, M. S. / Nobre-Oliveira, D. 2007. Effects of perceptual training on the iden-
tification and production of English voiceless plosives aspiration by Brazili-
an EFL learners. In Rauber, A. S. / Watkins, M. A. / Baptista, B. O. Eds.
New Sounds 2007. Proceedings of the Fifth International Symposium on the
Acquisition of Second Language Speech. Florianópolis: Universidade Feder-
al de Santa Catarina, 398–407. http://www.nupffale.ufsc.br/newsounds/
Papers/35.Reis_Nobre-Oliveira.pdf (2019-12-06).
Rost, C. A. 2008. A identidade do teuto-brasileiro na região sul do Brasil. Interdis-
ciplinar. Revista de Estudos em Língua e Literatura 5, 215–234. http://www.
seer.ufs.br/index.php/interdisciplinar/article/view/1125/963 (2019-12-06).
Ruiz Moreno, M. 2019. Phonetic production in early and late German-Spanish bi-
linguals. PhD dissertation. Hamburg: University of Hamburg.
Schaeffer, S. C. B. 2010. Oralidade e escrita de descendentes germânicos. Troca
de fonemas e de letras relacionadas à sonoridade. Cadernos do CNLF (Con-
gresso Nacional de Linguística e Filologia) 14, 1424–1443. http://www.
filologia.org.br/xiv_cnlf/tomo_2/1424-1443.pdf (2019-12-06).
Schaeffer, S. C. B. / Meireles, A. R. 2011. Padrões de vozeamento de consoantes
plosivas em falantes de pomerano (L1) e de português (L2). Anais do VII
Congresso Internacional da Abralin, Curitiba, 4009–4021. https://www.
academia.edu/2116193/ (2019-12-06).
Schaeffer, S. C. B. / Meireles, A. R. 2014. Descrição sonora da língua pomerana.
Letras de hoje 49, 46–55. http://revistaseletronicas.pucrs.br/ojs/index.php/
fale/article/viewFile/14700/11122 (2019-12-06).
Voice Onset Time in Pomerano-Brazilian Portuguese bilinguals
77
da Silveira, A. P. K. 2010. Escolas bilíngues em região de imigração. O caso de
Pomerode / SC. Revista da Associação Brasileira de Linguística 9, 41–71.
https://revistas.ufpr.br/abralin/article/view/52338/32237 (2019-12-06).
Stock, D. 1971. Untersuchungen zur Stimmhaftigkeit hochdeutscher Phonemrea-
lisationen. Hamburg: Buske.
Struck, E. 1992. Mittelpunktsiedlungen in Brasilien. Entwicklung und Struktur in
drei Siedlungsräumen Espírito Santos. Passau: Passavia.
Swartz, B. L. 1992. Gender difference in Voice Onset Time. Perceptual and Mo-
tor Skills 75, 983–992.
Tressmann, I. 1998. Bilinguismo no Brasil. O caso da comunidade pomerana de
Laranja da Terra/ES. Rio de Janeiro: Associação de Estudos da Linguagem
do Rio de Janeiro/Universidade Federal do Rio de Janeiro.
Tressmann, I. 2006. Dicionário enciclopédico Pomerano-Português. Pomerisch-
Portugijsisch Wöirbauk. Santa Maria de Jetibá: Secretaria de Educação.
Tressmann, I. 2007. O uso da língua no cotidiano e o bilinguismo entre pomera-
nos. Paper presented at the V Congresso Internacional da Associação Brasi-
leira de Linguística. Belo Horizonte, Brazil, 28 Feb –3 March 2007.
Tressmann, I. 2011. A educação escolar nas comunidades pomeranas. Ensino da
língua ou na própria língua? Revista da FARESE 3, 23–37.
Trudgill, P. 2000. Sociolinguistics. An introduction to language and society.
London: Penguin.
Vandresen, P. / Rodrigues Corrêa, A. 2008. O bilingüismo pomerano-português
na região de Pelotas. In Roncarati, C. / Abraçado, J. Eds. Português brasilei-
ro II. Contato linguístico, heterogeneidade e história. Niterói: Editora da
Universidade Federal Fluminense. http://www.leffa.pro.br/tela4/Textos/Textos
/Anais/CELSUL_VII/dir1/11.pdf (2019-12-06).
van de Weijer, J. / Kupisch, T. 2015. Voice Onset Time in heritage speakers and
second language speakers of German. In Babatsouli, E. / Ingram, D. Eds.
Proceedings of the International Symposium on Monolingual and Bilingual
Speech. Chania: Institute of Monolingual and Bilingual Speech, 414–420.
http://ismbs.eu/data/documents/Proceedings-ISMBS-2015.pdf (2019-12-06).
Wiese, R. 1996. The phonology of German. Oxford: Oxford University Press.
Willems, E. 1946. A aculturação dos alemães no Brasil. Estudo antropológico
dos imigrantes alemães e seus descendentes no Brasil. São Paulo: Compa-
nhia Editora Nacional.
Zampini, M. L. 2014. Voice Onset Time in second language Spanish. In Geeslin,
K. L. Ed. The Handbook of Spanish second language acquisition. Malden,
Wiley-Blackwell, 113–129.
Acknowledgments. We would like to express our gratitude to all the speakers who par-
ticipated in the study and to Edson Klemann, local activist for the preservation of
Pomerano culture, who kindly helped us to contact some of these speakers.
