=Paper=
{{Paper
|id=Vol-3128/paper12
|storemode=property
|title=Exploring consonant frequency in Sri Lanka Portuguese
|pdfUrl=https://ceur-ws.org/Vol-3128/paper12.pdf
|volume=Vol-3128
|authors=Carlos Silva,Luís Trigo
|dblpUrl=https://dblp.org/rec/conf/dhandnlp-ws/SilvaT22
}}
==Exploring consonant frequency in Sri Lanka Portuguese==
https://ceur-ws.org/Vol-3128/paper12.pdf
Exploring consonant frequency in Sri Lanka
Portuguese
Carlos Silva1⋆[0000−0002−8052−4271] and Luís Trigo2[0000−0002−3772−7081]
1
CLUP Centro de Linguística, University of Porto
2
LIACC Laboratório de Inteligência Artificial e Ciência de Computadores, University
of Porto
silvacarlosrogerio@gmail.com, trigoslab@gmail.com
Abstract. Although phoneme selection is a well-studied subject in con-
tact linguistics, phoneme integration is mostly unexplored. This study
aims at assessing phoneme integration by measuring consonant frequency
in Sri Lanka Portuguese and Portuguese. For that, we select two large lex-
ical corpora and, take several preparation steps to make the data uniform,
consistent and reusable. In terms of integration, we find that the more un-
constrained a consonant is concerning its phonotactic patterns, the more
frequent it is. We also find that being coronal has a positive impact on
integration, whereas being palatal has a negative impact. Moreover, we
find that in spite of the apparently random changes in the consonant fre-
quency, consonant classes are robustly transmitted from the lexifier to
this creole.
Keywords: Phoneme frequency · Sri Lanka Portuguese · Language con-
tact.
1 Introduction
Language contact leads to borrowing events between the languages involved.
On phonological grounds, these events often result in the addition of new phonemes
to the language’s inventory. Over the last decades, many researchers tried to
identify which segments were more or less more/less likely to be borrowed ei-
ther in individual cases (e.g. [9]) or more generally in the world’s languages
[13]. These studies give us a good idea of which segments are selected and why,
but do not report on how well integrated the phonemes are. Creoles make an
especially interesting case study, because they go beyond borrowing. They are
assumed to be a case of language shift [23].
While social motivations and structural compatibility are known to play a
role in phoneme selection during creole formation, the importance of the unin-
tentional psycholinguistic mechanisms is still to be explored. Usage frequency
⋆
Copyright © 2022 for this paper by its authors. Use permitted under Creative Com-
mons License Attribution 4.0 International (CC BY 4.0).
2 C. Silva & L. Trigo
is one of those mechanisms, which is crucial for the development of an individ-
ual’s linguistic knowledge [20,11]. Such factors are difficult to evaluate, espe-
cially when it comes to understudied languages, due to the lack of consistent
data or quantitative methods. However, thanks to recent developments, we are
finally able to start filling this gap.
This study aims at assessing the viability of this approach in contact linguis-
tics, as suggested by [14], through the exploration of consonant frequency in
two lexical corpora in a lexifier-creole pair, i.e., Portuguese and Sri Lanka Por-
tuguese. Our goal is to check the similarity rates between consonant frequency
in the lexifier and the creole, and to establish a possible link between the latter
and cross-linguistic frequency [19]. Thus, we expect to contribute to one as-
pect of the “creole debate”, by answering the question “Are the creoles similar
to their lexifier [8], their substrates [1] or do they reflect universal preferences
[3]?”.
Sri Lanka Portuguese can be seen as a typical case of “light creole” [6] or
a “trade creole” [2]. It was formed during the first half of the 16 th century,
after the establishment of a trading post at Colombo in 1517, born from the
contact between Portuguese, Sinhalese, and Tamil [22]. Later, it come into con-
tact with Dutch and English, which are considered its adstrates. The speakers
were members of the burgher communities, mainly represented by descendants
of Portuguese (and Dutch) men who married local women. After the English
took power over the island, this creole declined rapidly [10]. Nowadays, this
language is severely endangered, and it is spoken just by few members of the
Batticaloa and Trincomalee communities in the Tamil dominant east coast [16].
2 Methods
Although measuring consonant frequency is mathematically simple, the prepa-
ration steps are often challenging, especially for understudied languages. On
the one hand, if a relevant measurement should rely on a large lexical or usage
corpora with phonetic transcription, on the other hand, such corpora are rarely
available. Therefore, the first challenge is precisely converting orthographic
characters into IPA3 symbols.
As Sri Lanka Portuguese had no available corpus, we based our research on
the most extensive dictionary of this language [10], converted it into a con-
sistent format, and created PtLanka [24], an open-access online database that
assembles a total of 2522 words of Sri Lanka Portuguese. Then, we used a map-
ping table to assign phonetic symbols to the orthographic characters, according
to PHOIBLE conventions [19]. This step was taken to facilitate the future con-
version of the data into CLDF [12]. Finally, we split the phonetic symbols in
order to count them and assign them percentage values. Overall, this data set
is composed of 9 126 consonants.
3
IPA - International Phonetic Alphabet
Exploring consonant frequency in Sri Lanka Portuguese 3
For Portuguese, we extracted the entries of the Portuguese version of Wik-
tionary 4 . As a multilingual collaborative web-based project with phonetic in-
formation, wiktionary [18] seems to be a suitable tool, which does not imply
complex preparatory steps. However, as its structure relies on collective user
choices, it has no fixed structure. Therefore, before counting and assigning per-
centage values to the consonants, some preparation was needed. We started
by retrieving the IPA entries and cleaning irrelevant characters (e.g. =, ., *).
Then, we corrected the phonetic transcription, and convert some misplaced
SAMPA symbols into IPA. Finally, we group together some phonetic variants
into broader phonetic symbols (e.g. ɫ→l and ʁ→ʀ). This second data set has a
total of 49 674 consonants.
The method was developed with the Python programming language. The
implementation in Jupyter Notebook is available in the same repository as Pt-
Lanka with the aim of enabling transparency and reproducibility.
3 Results & Discussion
According to Bybee [4,5], the mental representations of language are inter-
nal representations of an individual’s experience. Then, if we can measure the
amount of an individual’s or a community’s experience with a given phoneme,
we can assess how well it is integrated in a given phonological system.
Table 1 shows the raw counts and the percentage values of each consonant
in Sri Lanka Portuguese (table 1a) and its lexifier (table 1b).
Regarding individual consonants, we can inspect the results in two dimen-
sions: integration and transmission. Integration corresponds to the degree to
which a consonant is frequent and, therefore, more strongly represented in
each individual language. Transmission correlates with the extent to which fre-
quency values are kept when comparing the creole and its lexifier.
If we look at PtLanka, we notice that there is no particular natural class on
the top, that is, /ɾ s d n/ do not share the same manner of articulation. However,
/ɾ s n/ are consonants that are allowed to occupy several syllable positions, such
as the onset (simple or complex) and the coda. Consequently, our data suggests
that the constrains on phonotactics influence frequency [17] and, therefore,
phonological integration. Furthermore, it is worth remarking that the seven
more frequent consonants are all coronal consonants. This finding corroborates
Carvalho’s proposal [7] which says that [coronal] is the unmarked point of
articulation within the oral cavity. On the contrary, palatal consonants /d̠ʒ t ̠ʃ
ɲ ʎ j ʃ lʲ nʲ/ are grouped on the bottom of the table 1a. This observation can be
explained, by their low frequency in the lexifier [25], on the one hand, and, on
the other hand, it confirms its high complexity [27,26,21].
When looking at both tables, we find no clues for a robust transmission at
a first glance. From those consonants with more statistical relevance, only /b g
z f/ have a range of less than 1% between the creole and the lexifier’s lexicon.
4
https://pt.wiktionary.org
4 C. Silva & L. Trigo
Table 1: Consonant frequency in Sri Lanka creole and Portuguese
(a) PtLanka
(b) Wikcionario
IPA count %
ɾ 1273 13,95 IPA count %
s 1055 11,56 ɾ 8132 16,37
d 905 9,92 t 5546 11,16
n 874 9,58 d 3741 7,53
t 847 9,28 l 3580 7,21
m 679 7,44 s 3537 7,12
p 545 5,97 k 3525 7,1
k 530 5,81 p 2525 5,08
l 461 5,05 m 2500 5,03
ʋ 310 3,4 ʃ 2321 4,67
b 308 3,37 n 1924 3,87
z 305 3,34 v 1486 2,99
f 246 2,7 b 1479 2,98
g 224 2,45 j 1364 2,75
r 160 1,75 w 1316 2,65
d̠ʒ 158 1,73 f 1315 2,65
t ̠ʃ 69 0,76 z 1224 2,46
ɲ 54 0,59 ʒ 1199 2,41
ʎ 42 0,46 g 1193 2,4
j 35 0,38 ʀ 1121 2,26
tː 16 0,18 ʎ 298 0,6
ʃ 15 0,16 ɲ 208 0,42
lʲ 13 0,14 kʷ 140 0,28
nʲ 2 0,02
Nevertheless, if we group the consonants into natural classes (e.g. stops, frica-
tives, etc), following [15], we recognize a robust transmission for most cases.
For instance, stops represent 36.5% of the phonemes in Sri Lanka Portuguese
and 36% in the lexifier language. The main positive difference is in the nasals,
whose usage is increased by 5.3% in the creole. On the opposite side, rothics
usage decreases about 3 percentual points. In light of these results, we conclude
that, whereas individual consonants may not show strong correlates between
the lexifier and the creoles, consonant classes do. The language contact itself
and historical change both in the Portuguese and Sri Lanka Portuguese may
have affected the phonetic shape of the segments but it didn’t have major ef-
fects on phonological classes as a whole. Thus, consonants in creole are not
simplified, they are adapted.
Although we believe to have reached some interesting results, this study
is far from complete. Firstly, we would like to condition the frequency of the
consonants and measure it in particular syllable positions (e.g. onset only). On
the one hand, this would make the consonants more comparable between them
Exploring consonant frequency in Sri Lanka Portuguese 5
and, on the other hand, it would serve as a test for the hypothesis above, i.e.,
phonotactics affects consonant frequency. In the second place, a comparison
with cross-linguistic frequency and cross-linguistic borrowability rates would
shed light on the influence of universal tendencies on phonological integration
in creoles. Finally, it would also be worth looking into other creoles which have
different languages as substrates and different contact situations, which would
complete our view on phonological integration and also bring some valuable
outcomes for historical linguistics and second language acquisition.
Acknowledgments This research was conducted within the doctoral program
of Languages Sciences (Faculty of Arts-University of Porto), was funded by the
Portuguese Foundation for Science and Technology (FCT MCTES) through the
PhD grant SFRH/BD/2020.07466.BD and supported by the Center of Linguistics
of the University of Porto (FCT-UIDB/00022/2020).
References
1. Alleyne, M.: Comparative Afro-American: an historical-comparative study of
English-based Afro-American dialects of the New World. Karoma, Ann Arbor (1980)
2. Bakker, P., Daval-Markussen, A., Plag, I., Parkvall, M.: Creoles are typologically dis-
tinct from non-creoles. Journal of Pidgin and Creole Languages 26(1), 5–42 (2011).
https://doi.org/10.1075/jpcl.26.1.02bak
3. Bickerton, D.: Roots of Language. Karoma (1981)
4. Bybee, J.: Phonology and Language Use. Cambridge Studies
in Linguistics, Cambridge University Press, Cambridge (2001).
https://doi.org/10.1017/CBO9780511612886
5. Bybee, J.: Language, Usage and Cognition. Cambridge University Press, Cambridge
(2010). https://doi.org/10.1017/CBO9780511750526
6. Carvalho, A., Lucchesi, D.: Portuguese in contact. In: The Handbook
of Portuguese Linguistics, pp. 41–55. Wiley Blackwell (apr 2016).
https://doi.org/10.1002/9781118791844.ch3
7. Carvalho, J.B.D.: Why there is no backness: the case for dismissing both [coronal]
and [dorsal]. In: Naïm, J.L.L..S. (ed.) Backness and backing, pp. 45–58. Lincom
(2013), https://halshs.archives-ouvertes.fr/halshs-01116259
8. Chaudenson, R.: Des îles, des hommes, des langues: essai sur la créolisation linguis-
tique et culturelle. L’Harmattan (1992)
9. Clements, J.: The status of portuguese/spanish /r/ and /r/ in some iberian-based
creole languages. PAPIA 24(2), 343–356 (2014), http://revistas.fflch.usp.br/
papia/article/view/2201
10. Dalgado, S.R.: Dialecto indo-português de Ceylão. Imprensa Nacional, Lisboa (1900)
11. Edwards, J., Beckman, M., Munson, B.: Frequency effects in phono-
logical acquisition. Journal of child language 42, 306–11 (03 2015).
https://doi.org/10.1017/S0305000914000634
12. Forkel, R., List, J.M., Greenhill, S., Rzymski, C., Bank, S., Cysouw, M., Hammarstrom,
H., Haspelmath, M., Kaiping, G., Gray, R.: Cross-linguistic data formats, advanc-
ing data sharing and re-use in comparative linguistics. Scientific Data 5 (2018).
https://doi.org/10.1038/sdata.2018.205, https://doi.org/10.1038/sdata.2018.
20
6 C. Silva & L. Trigo
13. Grossman, E., Eisen, E., Nikolaev, D., Moran, S.: Segbo: A database of borrowed
sounds in the world’s languages. In: Proceedings of the Twelfth International Con-
ference on Language Resources and Evaluation (LREC 2020) (2020), http://www.
lrec-conf.org/proceedings/lrec2020/pdf/2020.lrec-1.654.pdf
14. Hakimov, N., Backus, A.: Usage-based contact linguistics: Effects of frequency and
similarity in language contact. Journal of Language Contact 13(3), 459 – 481 (2021).
https://doi.org/https://doi.org/10.1163/19552629-13030009
15. Ladefoged, P., Maddieson, I.: The Sounds of the World’s Languages. Wiley (1996)
16. Lee, N.H.: The status of endangered contact languages of the
world. Annual Review of Linguistics 6(1), 301–318 (2020).
https://doi.org/10.1146/annurev-linguistics-011619-030427, https://doi.
org/10.1146/annurev-linguistics-011619-030427
17. Macklin-Cordes, J., Round, E.: Re-evaluating phoneme frequencies. Frontiers in
Psychology 11 (2020). https://doi.org/10.3389/fpsyg.2020.570895, https://www.
frontiersin.org/article/10.3389/fpsyg.2020.570895
18. Meyer, C.M., Gurevych, I.: Wiktionary: A new rival for expert-built lexicons? Ex-
ploring the possibilities of collaborative lexicography. na (2012)
19. Moran, S., McCloy, D. (eds.): PHOIBLE 2.0. Max Planck Institute for the Science of
Human History, Jena (2019), https://phoible.org/
20. Munson, B.: Phonological pattern frequency and speech production in adults and
children. Journal of speech, language, and hearing research 44, 778–92 (09 2001).
https://doi.org/10.1044/1092-4388(2001/061)
21. Silva, C.: The representation of portuguese palatal sonorants through the eyes of
portuguese-based creoles. In: Proceedings of the 8th School-Conference Language
issues: a young scholars’ perspective. Institute of Linguistics of the Russian Academy
of Sciences, Moscow (forc)
22. Smith, I.R.: Sri lanka portuguese structure dataset. In: Michaelis, S.M., Maurer, P.,
Haspelmath, M., Huber, M. (eds.) Atlas of Pidgin and Creole Language Structures
Online. Max Planck Institute for Evolutionary Anthropology, Leipzig (2013), https:
//apics-online.info/contributions/41
23. Thomason, S.G., Kaufman, T.: Language contact, creolization, and genetic linguis-
tics. University of California Press (1988)
24. Trigo, L., Silva, C.: Ptlanka: an online corpus of sri lanka portuguese lexicon and
phonology. In: OpenCor 2021 (2021)
25. Trigo, L., Silva, C.: Comparing lexical and usage frequencies of palatal segments in
portuguese. In: Proceedings of the 15th edition of the International Conference on
the Computational Processing of Portuguese (PROPOR 2022). Springer, Fortaleza
(forc)
26. Veloso, J.: Complex segments in portuguese: The unbearable heaviness of be-
ing palatal. In: Zendoia, I.E., Nazabal, O.J. (eds.) Bihotz ahots. M. L. Oñederra
irakaslearen omenez, pp. 513–526. Euskal Herriko Unibertsitatea, The address of
the publisher (2019)
27. Wetzels, W.L.: Consoantes palatais como geminadas fonológicas no por-
tuguês brasileiro. Revista de Estudos da Linguagem 9(2), 5–15 (2000).
https://doi.org/10.17851/2237-2083.9.2.5-15, http://www.periodicos.letras.
ufmg.br/index.php/relin/article/view/2323