RECONSTRUCTING KRIOL SYLLABLE STRUCTURES (original) (raw)
retornar ao índice dos papers|retornar à pagina do LIV | retornar à página da UnB
Reconstructing Kriol syllable structures*
Hildo Honório do Couto
(Universidade de Brasíla)
The Portuguese pidgin/creole (PPC) formed along the African coast in the 15th, 16th and 17th centuries is of paramount importance for the history of creolistics. See, for instance, the monogenetic theory presented by many of its historians to explain the origin of several creoles around the world. However, this importance has not been reflected in the number of studies devoted to PPC. The English and French-based creoles are much better studied, both structurally and historically.
To my knowledge, the first time the word "crioulo" was even used in reference to a PPC was in 16841. The first creole phrase was recorded as late as 1696, in a rather Lusocentric way. But we had to wait until the end of the 19th century to see texts written in this creole2. As in the case of the phrase just mentioned, they were written in a Lusitanized way. The only documents we have from the 15th to the 19th centuries are some isolated words whose morphological - not to mention their syntactic behavior - is entirely unknown.
Due to this lack opf records in the form of texts and complex words (compound and derived) - if they ever existed - from the period of PPC formation, it is difficult, if not impossible, even to try to reconstruct its syntax or its morphology. Thus, the aim of this paper is to show that it is possible to give at least some hints as to what the syllabic structures of Kriol - the continental variant of PPC - may have been. I shall then go on to show some of the transformations that the postulated syllable structures of the formative period may have undergone. In this period there was a_dephonologization_3 of Portuguese structures, followed by a phonologization, i.e., a structuring of a new language. What happened later was due to a _rephonologization_of Kriol structures in the direction of the superstrate language, namely, Portuguese. As we can see, the three concepts - independently proposed by Jakobson in 1939 - are the rough equivalents of, respectively,pidginization, creolization and decreolization (see Jakobson 1978).
To begin with, I resorted to a few isolated words recorded in the period from 1506 to 1696 (see Appendix!). Assuming that they were pronounced approximately as they are in today’s basilect4, we may set up the consonant and vowel systems seen in (1)(a)-(b), where c = tch (like "ch" in English spelling), j = dj (like "j" in "judge") and N stands for the nasal velar (like the "ng" of English).
(1) (a) (b)
p t c k i u
b d j g e o
m n ñ N a
f s
l
r
w y
Swadesh’s list of 100 basic words corrected by Cassidy (1971) leads to the same results. In Mbodj (1979: 46, 48) we have the same phonemes for the language of today, except for /w, y/5. However, as shown in Couto (1994), today’s Kriol additionally possesses the following consonants, borrowed from Portuguese: (i) voiceless palatal fricative (spelled "sh"in English), (ii) voiced palatal fricative (spelled "j" in French), (iii) palatal lateral (representd by the Greek letter lambda). In other words, it has practically all the phonemes of the acrolect, but that is due to the rephonologization of Kriol structures in the direction of Portuguese, i.e., decreolization.
The list of words of the 1506-1696 period enables us to postulate the syllable structures for PPC in (2), given with percentage of occurrence. This table differs somewhat from the one obtained from Swadesh/Cassidy’s list of 100 words, shown in (3).
(2) (3)
CV 75.23% CV 68.89%
CVC 18.15 CVC 22.00
CCV 3.33 V 4.30
V 2.38 VC 1.91
CCVC 0.47 CCV 1.91
CCVC 0.47
CCCV 0.47
On the other hand, the latter table (3) corresponds exactly to the tables arrived at in Couto (1994) and Mbodj (1979: 54), except for the CCCV pattern, which is handled separately by both authors. In (4) we have some instantiations of this structure given by the first, but not by the second, author.
(4) strela (< Port. "estrela") ‘star’
strada (< Port. "estrada") ‘way’
splika (< Port. "explicar") ‘to explain’
skribi (< Port. "escrever") ‘to write’
Like the additional phonemes seen above, the forms in (4) tend to occur only in an extremely decreolized variety of Kriol, as we shall see below.
According to Clements & Keyser (1983: 28), the most unmarked syllable structure is CV, "may be operated on to yield one or more of the other core syllable types by the following operations":
(5)
- delete syllable inicial C
- insert syllable final C
The authors continue: "Any language may choose either, both, or neither of these two rules to expand its inventory of primary core syllables". In (6) we have the language types arrived at by setting the parameters of (5) - in parentheses we have the example of one language of this type.
(6)
Type I: CV (Senoufo)
Type II: CV, V (Maori)
Type III: CV, CVC (Klamath)
type IV: CV, V, CVC (English)
(Clements & Keyser 1983: 29)
From top to bottom we go from the least marked to the most marked language types as regards syllable structure.
Now, if we turn to (3) or even to (2) above, we shall note that Kriol would be a type IV language, i.e., it shows the most marked syllable strutures possible by fixing the parameters of (5). And what is more, it selects among some further options to generate the syllable structures like the CCCV pattern of (3). That runs contrary to the common assumption that creole structures are less marked than non-creole language structures, specifically those of the superstrate language. Portuguese dos not have underlying CCCV structures.
Let us begin by examining the structures shown in (2). As already noted, they were obtained from the words recorded in the period of PPC formation, under the assumption that they were pronounced more or less as in the more conservative varieties of Kriol. Complex syllable structures like CCV and CCVC could be represented by words like "prasa" (Port. "praça") ‘town’ and "kriston" (Port. "cristão") ‘Christian’. If we take into consideration the fact that the few words that were recorded up to the 19th century were seen from a Lusocentric perspective - even today writers give "papiá" cristão" for Papia Kristang -, it is legitimate to hypothesize that "prasa" and "kriston" were not the most common pronunciations of these words. It is, thus, highly probable that the first syllable of "kriston" and "prasa" had an epenthetic vowel, at least in the period immediately following the dephonologization of Portuguese structures, introduced by speakers of substrate languages. In other words, by the time the early Portuguese pidgin had stabilized and/or the phonologization of PPC structures had taken place, these words may have been pronounced [parasa] and [kiriston], respectively.
There are many synchronic evidences that support the epenthesis hypothesis. One of them is represented by the phonologization of Portuguese words like those in (7).
(7)
garandi (<Port. "grande") 'big'
kiriol (<Port. "crioulo") 'creole'
kuru (<Port. "cru") 'raw, uncooked'
birinka (<Port. "brincar") 'to play'
firiu (<Port. "frio") 'cold'
garan (<Port. "grão") 'grain, corn'
The third and fourth examples in (7) were taken from Chataigner (1963: 48). As is well known, the Casamance dialect he studied is a more conservative variety of Kriol due to the loss of contact with the substrate language as early as 1886. Besides that, it is a geographic continuation of Cachéu creole. One of the most deep-rooted beliefs among Kriol speakers in Guinea-Bissau is that it originated in this town and in Geba.
A second synchronic evidence for syllable simplification, i.e., syllable change towards CV, is represented by the examples in (8).
(8)
tarbaja (<Port. "trabalhar") 'to work'
gardisi (<Port. "agradecer") 'to thank'
ferga (<Port. "esfregar") 'to rub'
pirgisa (< Port. "preguiça") 'laziness'
Although CVC is more complex than CV, it is, nevertheless, far less marked than CCV according to the principles presented in Clements & Keyser (1983). In other words, it results directly from the application of rule (5b) - one of the initial parameters - as is the case with CCV.
Several other changes that Portuguese words underwent in the early period of PPC formation point towards CV syllable structure. Further examples are presented in (9).
(9)
tisi (< Port. "trazer") 'to bring'
pati (< Port. "repartir") 'to give, share out'
tona (<Port. "tornar") 'to return, to do again'
talabes (<Port. "talvez") 'maybe, perhaps'
In summary, in its formative period Kriol may have avoided applying not only (5b) but even (5 a). Since V syllable structures were also avoided, as seen in the examples of (10), we can conclude that the early PPC had a preference for CV syllables.
(10)
kil (<Port. "aquele") 'that, this'
raña (<Port. "aranha") 'spider'
lifanti (<Port. "elefante") 'elephant'
gora (<Port. "agora") 'now'
PPC tended to be a type I language. This claim is independently put forward by Chataigner who says: "Mais il reste que le kriol connaît d'autres types de mots, où la syllabe n'est plus du type CV pur. Les examinant de près, il semble pourtant qu'on puisse en réduire beaucoup à CV, si l'on adopte de C une conception africaine" (Chataigner 1963: 48). Obviously, his extreme substratist point of view is not wholly correct but I think he points in the right direction.
What later happened to Kriol syllables, providing structures like those in (4), was the well-known process of complexification that always takes place in the transition from a pidgin to a creole (Hymes 1971: 77, 84) and, especially, in decreolization. In Jakobson's phonological terms, it was the phonologization and rephonologization of the initially dephonologized Portuguese phonetic material.
The structures exemplified in (4) above are the most marked that we can find in Kriol, i.e., CCCV. Since one of the words that show this structure is included in Swadesh/Cassidy's list of 100 basic words, one could conclude that the above hypothesis about the tendency of PPC towards CV is falsified. However, we should note, first, that Clements & Keyser (1983) say that "any three-member cluster, ABC, is analysable into two-member clusters, AB and BC, each of which safisfies"the syllable structure conditions (SSC) of the language in question. At least in its later stages, Kriol possessed the SSCs shown in (11 a-b). Thus, the /pr, pl, tr, kr, kl/ clusters - exemplified in the lisf of word of the 1506-1696 period - are generated by (11 b), whereas the /sp, st, sk/ clusters - exemplified by "sta" (to be, to stand) - are generated by (11 a). I is important to remark that the SSCs in (11) apply if we ignore the simplification that is exemplified in (7) above. It must have applied very late in the history of Kriol, in any case.
(11) (a) O (b) O
/ \ / \
/ \ / \
C1 C2 C1 C2
[+strid] [-cont] [-son] [+son]
[+cor] [+s.v.c.] [-nas]
(11 a) says that in an onset having a strident coronal as the first consonant (C1), the secont consonant (C2) may be any obstruent ([-cont]) with stiff vocal cords ([+s.v.c.]), namely, /p, t, k/. (11 b) shows that evey onset which has an obstruent ([-son]) as C1 must have a non-nasal sonorant - the so-called liquids - as C2.
In the secont place, even leaving aside the above observation that the transcription of the few recorded words was very Lusocentric, "sta" has the variant "ta" in some dialects of Kriol. In Capeverdian it has, additionally, the variant "sata". In the third place - and this is the most decisive evidence -, the majority of the exemples in (4) have a simpler synchronic alternant. Thus, "splika" appears also under the form "siplika", and "skribi" is also used under the form "skirbi"(see Couto 1994). According to the principles seen above, CCVC (skir-) is less marked than CCCV (skri-). In "siplika", we have even the less marked syllable initially, followed by a CCV structure. That is to say, if the words in (4) were used at all in the period of dephonologization of Portuguese syllabes and of phonologization of PPC, it is very likely that they tended to conform to a language of type III or even of type I, in the latter case by applying neither of the (5 a) nor (5 b). To put it differently, at that time the language tended to have only the most general - less marked - CV syllable structure.
A well-known fact about today's Kriol is the application of rule (5 b), hence the generation of the syllable structures CVC and VC. Words having syllables like those in (12) are very common today (syllables are separated by a dot).
(12)
kri.ol / ki.ri.ol 'creole, Kriol
kar.nel (<Port. "carneiro) 'sheep'
a .os (<Port. "hoje") 'today'
os (< Port. "osso") 'bone'
Additional synchronic evidence in favor of the thesis of the initial tendency towards CV is given by today's variation like those in (13). The simpler form, i.e., the one having CV structure, is more conservative, whereas the more complex one is innovative.
(13)
jukuta / jukta 'to jump'
sukuta / sukta 'to hear'
jumuna / jumna 'to arrive first'
But following all the above observations, this is certainly a late addition to the original system - perhaps earlier than the addition of a secont initial C, as envisaged by Clements & Keyser's theory. At any event, VC structures may have been in use as early as the end of the 15th century.
To summarize, in its formative period, PPC tended to be a language of type I (CV). But, as soon as the early unstable pidgin began to stabilize and became Kriol, its syllable structures tended to become more complex due to the increase in its use and, presumably, to the influence of the superstrate - perhaps due to the substrate languages as well. This move led Kriol in the direction of type III and, eventually, of type IV. In its later period of evolution, the language adopted the structures shown in (4) by the combination of SSCs (11) (a) and (b).
The above conclusions are in accordance with Bickerton's hypothesis that claims that in the creation of a creole by the first generation of pidgin speaker's children, the universal principles of the language bioprogram tend to appear in all their strength (Bickerton 1984). What happened later was due to the well-known complication process that is formally envisaged by Clements & Keyser's syllable structure theory. The scarce examples recorded from 1506 to 1696 (see Appendix!) give these assumptions some empirical support.
NOTES
*This is a revised version of a paper read at the III Meeting of the Society for Pidgin and Creole Linguistics, Philadelphia, January 9-12, 1992. Later it was presented at the Encontro Nacional da ANPOLL, Porto Alegre, 17-20/5/1992. There's a copy of it in the Anais do VII Encontro Nacional da ANPOLL vol. 2, Goiânia, 1993, p. 600-608.
1. For this and other historical references, see Couto (1993, 1994).
2. See Couto (1994, especially sections 1.5 and 4.0.
3. For the unstructuredness of an early pidgin, see Bickerton (1984: 174).
4. There is not enough space here to justify this assumption, but I have some evidences that it is correct. I intend to explore them some day.
5. For the phonologization of Portuguese sounds into PPC, see Couto (1994, sections 3.1.1 through 3.1.3).
REFERENCES
Bikerton, Derek. 1984. The language bioprogram hypothesis. In: The behavioral and brain sciences 7.173-221.
Cassidy, F. G. 1971. Tracing the pidgin element in Jamaican creole. In: Hymes, D. (ed.) Pidginization and creolization of languages. Cambridge: Cambridge University Press: 203-221.
Chataigner, Abel. 1963. Le créole portugais du Sénégal: observations et textes. Journal of African languages 1,1.44-71.
Couto, Hildo Honório do. 1993. The genesis of Portuguese creole in Africa. In Holm, John & Frank Byrne (eds.). Amsterdam: Benjamins, p. 381-389. Atlantic meets Pacific: a global view of pidginization and creolization
_______. 1994. O crioulo português da Guiné-Bissau. Hamburg: Helmut Buske Verlag.
Jakobson, Roman. 1978. Principles of historical phonology. In: Baldi, P. & R. R. Werth (eds.) Readings in historical phonology. University Park and London: The Penn State University Press: 103-120.
Mbodj, Chérif. 1979. Phonologie du créole de Guinée-Bissau. Dakar: CLAD.
APPENDIX
adibi 'sword'
anima 'idol'
atagara 'wooden bowl'
bafatá 'Bafatá' (a town)
baga 'ethnic group'
bagri 'a kind of fish'
balafon 'a kind of musical instrument'
balanga 'ethnic group'
baloba 'idol's house'
banana 'banana'
bandin 'quarter of Bissau'
batanga 'rice candy'
becerin 'wizard'
bijago 'ethnic group'
biñu 'wine'
bisau 'Bissau'
bonbolon 'a kind of drum'
bufaru 'buffalo'
bugan 'place name'
Bulama 'Bolama' (a town)
buramu 'ethnic group'
daba 'hoe'
deponga 'a town'
di 'of'
cai 'adultery'
cina 'idol'
coka 'partridge'
colona 'interpreter'
ceben 'palm oil'
fanadu 'circumcision'
farin 'emperor'
fulu 'fula, peul'
ganagoga 'polyglot'
ginala 'place name'
gine 'Guinea'
goyaba 'guava'
grumeti/gurmeti/gurumeti 'grumetto'
iñami 'yam'
jagra 'caste'
jalofu 'Wolof'
janbakus 'witchdoctor'
janta 'lunch'
jidiu 'griot'
kabasera 'Adonsonia digitata'
kabu 'place'
kaciu 'Cachéu" (a town)
kalanbe/kalame 'virginity's cloth'
kola 'cola acuminata'
kriston 'Christian'
kuskus 'couscous (French)
lagargu 'crocodile'
lala 'open place to plant rice'
makariu 'sudden flood'
manduku 'cudgel'
mandimansa 'chief'
mandinga 'Mandingo' (ethnic group)
mankara 'peanut'
marabu 'marabou'
marlota 'Moorish clothes'
melegeta/malageta/malgeta 'Indian pepper'
manpata 'parinari excelsa'
negru 'black, negro'
nomina 'bijouterie'
palma 'palm'
pañun 'ethnic group'
patata 'potato'
polon 'Eriodendron anfractuosum'
prasa 'town'
sabi 'good, tasty'
siga 'antelope'
sta 'to be, to stand'
sumana 'week'
tangoma 'acculturated woman who acted as European men's concubine'
tapada 'fence'
tara 'type of palm tree'
ternadu/tornadu 'tornado'
trusiman 'interpreter'
retornar ao índice dos papers|retornar à pagina do LIV | retornar à página da UnB