Pidgins and Creoles in Asia

!"#$%&''()

Pidgins and Creoles in Asia

!"#$%&''()

Volume 38

Pidgins and Creoles in AsiaEdited by Umberto Ansaldo!ese materials were previously published in Journal of Pidgin and Creole Languages 25:1 (2010)

Benjamins Current TopicsSpecial issues of established journals tend to circulate within the orbit of the subscribers of those journals. For the Benjamins Current Topics series a number of special issues of various journals have been selected containing salient topics of research with the aim of finding new audiences for topically interesting material, bringing such material to a wider readership in book format.

For an overview of all books published in this series, please see http://benjamins.com/catalog/bct

Pidgins and Creoles in Asia

Edited by

Umberto Ansaldo!e University of Hong Kong

John Benjamins Publishing CompanyAmsterdam / Philadelphia

!"#$%&''()

Library of Congress Cataloging-in-Publication Data

Pidgins and Creoles in Asia / edited by Umberto Ansaldo.p. cm. (Benjamins Current Topics, - ; v. )Includes bibliographical references and index.. Pidgin languages--Asia. . Creole dialects--Asia. . Languages in contact--Asia.

I. Ansaldo, Umberto.PM.AP ’.--dc (Hb ; alk. paper) (Eb)

© !"#! – John Benjamins B.V.No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher.

John Benjamins Publishing Co. · P.O. Box $%!!& · #"!" '( Amsterdam · )e NetherlandsJohn Benjamins North America · P.O. Box !*+#, · Philadelphia -. #,##/-"+#, · 01.

!e paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, z39.48-1984.

Table of contents

Foreword vUmberto Ansaldo

Chinese Pidgin Russian 1Roman Shapiro

China Coast Pidgin: Texts and contexts 59Umberto Ansaldo, Stephen Matthews, and Geo! Smith

!e African slave population of Portuguese India: Demographics and impact on Indo-Portuguese 91

Hugo C. Cardoso

Vestiges of etymological gender in Malacca Creole Portuguese 115Alan N. Baxter

Bazaar Malay topics 151Bao Zhiming and Khin Khin Aye

Index 169

!"#$%&''()

Foreword

Asia is a very large and diverse region of the globe, so much so that there are at times doubts whether it constitutes an entity at all that can be captured under one name (Lim & Low 2009). Certainly from the point of view of a linguist, the diversity of language families, the historical time depth, the complex patterns of population movements, and the wealth of contact phenomena that de"ne Asia are so many and at times still so little understood that no single volume could ever pretend to shed su#cient light on even one of these aspects of the region. !is volume— previously published as JPCL 25:1 — looks at some pidgin and creole varieties of Southeast Asia, Southern China, and the Russian-Chinese border re-gions; it is meant to be a sample platter, an amuse bouche of the "eld of contact linguistics in this part of the world (see also Lim & Gisborne 2009 on Asian vari-eties of English).

Traditionally strongly rooted in the study of Atlantic varieties, with the Paci"c also somewhat well represented, the "eld of pidgin and creole studies tends to feature Asian contexts of contact less prominently (but see Holm 1989; Wurm et al. 1996). Recently the phenomenon of Singlish, or Singapore Colloquial English, has come to the fore, thanks in particular to the work of Bao (2001, 2003, 2005, 2009) and Lim (2004, 2008, 2009) who, by focusing on actual basilectal/mesolectal Singlish — rather than English spoken by highly educated people in Singapore — reveal the many substrate features of what is really an English-lexi"er contact lan-guage with a strong Sinitic and possibly Malay substrate (Ansaldo 2004). Within Austronesian studies it has long been known that contact varieties of Malay have existed in the Malay/Indonesian region for probably over a millennium (Adelaar & Prentice 1996), but these languages have only recently become the focus of at-tention within contact linguistics (despite the interest already shown in Hugo Schuchardt’s work). I would have liked to include in this volume several contact-Malay varieties, such as Baba Malay and even Sri Lanka Malay, but in this instance I opted for just one, in order to represent as diverse a range of topics and varieties as possible (see Ansaldo 2009 for contact varieties of Malay in Monsoon Asia).

!is volume starts with what I see as a real treat: a state-of-the-art paper on Chinese Russian Pidgin, a variety that has received some attention in Russian but very little in English. In his paper Shapiro gives us a detailed account of the inter-play between Sinitic and Russian elements as well as, intriguingly, the Tungusic

!"#$%&''()

vi Foreword

and Mongol features that in$uenced the grammar of this pidgin as it evolved along the banks of the Amur River. Chinese was traditionally one of the big players in the region long before Western colonization, and is thus appropriately the focus of the second paper in this volume. Widely accepted to be the variety that gave the "eld the term ‘pidgin’, Chinese Pidgin English or China Coast Pidgin, as it is re-ferred to in Ansaldo, Matthews, & Smith’s paper, is a remarkable story thanks to its abundant written texts that yield data far more Sinitic in nature than so far gener-ally assumed. !e third and fourth papers are dedicated to the most in$uential of the colonial languages that accompanied early Western intrusion into the region: Portuguese. Cardoso focuses on Indo-Portuguese population history, shedding important light on the composition of the slave population in the Indian territo-ries and discussing its signi"cance for the formation of Indo-Portuguese varieties. Baxter moves eastward to Malayo-Portuguese, presenting data on in$ectional ma-terial in the Papia Kristang of Melaka which lead to an interesting discussion of residual morphology in creole-formation. !e "%h and last contribution is on the ubiquitous Bazaar Malay spoken in Singapore and, in di&erent varieties, in many other ports of the region. In their paper Bao & Khin Khin Aye describe topicaliza-tion in Bazaar Malay, revealing strong Sinitic substrate in$uences, and giving us a taste of this historically signi"cant Malay-Sinitic lingua franca of the region.

!is collection is intended to whet our appetites for a whole new range of typological admixtures, socio-historical settings, and de"nitional challenges. In Asian contexts one has the opportunity to observe contacts of di&erent kinds, in-cluding those between the ‘local’ languages, such as Sinitic and Malay, as well as those involving western/colonial languages (such as English and Portuguese) and local substrates. I believe that understanding Asian contexts and integrating the generalizations that we derive from their observation into our current theories of contact language formation is bound to enrich the "eld and make our theoretical constructs more solid and objective. I am very grateful to the editors of JPCL for giving me the opportunity and the generous support to make this volume a reality, as well as to the hard work of the authors and the reviewers to whom all the merit goes. Any shortcomings are solely my responsibility.

Umberto Ansaldo "e University of Hong Kong

September 2009

Foreword vii

References

Ansaldo, Umberto. 2004. !e evolution of Singapore English: Finding the matrix. In Lisa Lim (ed.), Singapore English: A grammatical description (Varieties of English Around the World G33), 127–149. Amsterdam/Philadelphia: John Benjamins.

Ansaldo, Umberto. 2009. Contact languages. Ecology and evolution in Asia. Cambridge: Cambridge University Press.

Bao, Zhiming. 2001. !e origins of empty categories in Singapore English. Journal of Pidgin and Creole Languages 16.2: 275–319.

Bao, Zhiming. 2005. !e aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics 41. 237–267.

Bao, Zhiming. 2009. One in Singapore. Studies in Language 33(2). 338–365.Bao, Zhiming & Lye Hui Min. 2005. Systemic transfer, topic prominence, and the bare condi-

tional in Singapore English. Journal of Pidgin and Creole Languages 20.2: 269–291.Holm, John. 1989. Pidgin and creole languages. Vol. II. Cambridge: Cambridge University Press.Lim, Lisa. 2004. Sounding Singaporean. In Lisa Lim (ed.), Singapore English: A grammatical

description (Varieties of English Around the World G33), 19–56. Amsterdam/Philadelphia: John Benjamins.

Lim, Lisa. 2007. Mergers and acquisitions: On the ages and origins of Singapore English par-ticles. World Englishes 27.4: 446–473.

Lim, Lisa. 2009. Revisiting English prosody: (Some) New Englishes as tone languages? In Lisa Lim & Nikolas Gisborne (eds.), "e typology of Asian Englishes, Special Issue, English World-Wide 30(2): 218–239.

Lim, Lisa & Nikolas Gisborne (eds.). 2009. "e Typology of Asian Englishes. Special Issue, English World-Wide 30(2).

Lim, Lisa & Ee-Ling Low (eds.). 2009. Multilingual, globalizing Asia: Implications for policy and education. Special Issue, AILA Review 22.

Schuchardt, Hugo. 1890. Kreolische Studien IX. Ueber das Malaioportugiesiche von Batavia und Tugu. Sitzungberichte der kaiserlichen Akademie der Wissenscha# zu Wien. 122(9): 1–256.

!"#$%&''()

Chinese Pidgin Russian*

Roman ShapiroRussian State University for the Humanities & University of Malaya

!e much-understudied Chinese Pidgin Russian (CPR) has existed at the Chinese–Russian border since at least the 18th century. Unlike many Western-based pidgins, it was formed in a territory where the lexifying language (Russian) was dominant. It also uses a typical in$ecting language as its lexi"er and an isolating language (Chinese) as its substrate. !is paper considers the in-$uence of both ‘parent’ languages at all CPR levels. !e sources of CPR include: pidgin records and descriptions; ‘Russian’ textbooks compiled for the Chinese going to Russia; and works of literature depicting contacts between the Russians and indigenous peoples of Siberia, who o%en spoke a variety of CPR. Some of these sources are rarely accessible to Western linguists. !e paper discusses all key aspects of CPR: history (both of the pidgin and its study), phonology (seg-mental inventory, stress, tone), morphology (verbs vs. non-verbs, "nal particles), syntax (syntactic roles, sentence and phrase word order, postpositions and prepositions, comparatives), and vocabulary (synonyms, loanwords, structural and semantic calques, ‘diminutive politeness’). !e study provides new transla-tions and etymologies for ‘di#cult’ CPR words and sentences.

1. Introduction

!e last few decades have witnessed a trend towards generalisation and search for common structural characteristics in pidgin studies. However, speci"c research is still focused on pidgins based on Western European languages, such as Portuguese, French, or English. In order to make accurate cross-linguistic generalisations, it is important to take into account contact languages that emerged in di&erent lin-guistic and social circumstances (for example, Ansaldo 2009 on Asian contexts, Siegel 2000 for the Paci"c, as well as Winford 2006, !omason & Kaufman 1998

* !anks are due to the Chiang Ching-kuo Foundation and the European Association of Chinese Studies for a Library Travel Grant, which was essential for preparing this paper. !e au-thor is also grateful to Dr. Umberto Ansaldo, Dr. Donald Winford, and an anonymous reviewer for their advice.

!"#$%&''()

2 Roman Shapiro

for a broad cross-linguistic framework). Unlike many Western-based pidgins, the formation of Russian pidgins did not involve slavery and took place in a territory where the lexi"er (Russian) was dominant. Russian is also a typical in$ecting lan-guage, so Russian pidgins allow seeing how this feature is treated in pidgins.

Russian-Norwegian pidgin (Russenorsk) and Taimyr Pidgin Russian (Govorka) have been described to some extent (Broch 1927, Broch 1930, Broch & Jahr 1981, 1984, 1990; and Xelimskij 1987, 1996, 2000, Stern 2002, 2005a). Other Russian-lexicon pidgins, including Chinese Pidgin Russian (CPR), have received until now relatively little attention.1

2. Sources and a brief history

Chinese-Russian contacts are "rst documented in the 13th century. !e travelogue of Jean Plano de Carpini (1182–1252) mentions Russians in China. Russian cap-tives are described in "e history of Yuan under years 1331–1332. "e secret history of the Mongols (13th century, a%er 1227) is the "rst Chinese source (originally in Mongolian) that mentions Russia.2 !e Russians "rst mention China as part of the Mongolian Empire in "e second Sophia chronicle under year 6903 (i.e. 1394–1395).3 More active contacts began in the 17th century, when Russians sent several missions to China (the "rst to reach China was Petelin’s mission, which arrived in Beijing in September 1618). Reasonably stable caravan trade began a%er 1679, when China and Russia signed a tea-trade treaty. For a survey of early Chinese Russian contacts see Yan (2006: 20–87), Skachkov (1966, 1977), and Demidova & Miasnikov (1966). Although we cannot rule out the existence of a Sino-Russian pidgin at that period, no relevant data have been available so far.

According to the treaty of Kyakhta (1727), Kyakhta in Russia and Maimachin (买卖城)4 in China were founded as special towns designated for border trade. A%er 1757, when caravan trade was abolished, and until 1860, when the treaty of Beijing li%ed the restrictions on the Russian-Chinese trade, Kyakhta/Maimachin

1. For a complete list of Russian-based pidgins see Stern 2002.

2. Incidentally, it was "rst discovered and published in the West by a Russian, Archimandrite Palladius (Kafarov), in 1866.

3. !e Mongolian Yuan dynasty was actually overthrown in China in 1368.

4. Maimachin is today’s Altanbulag (Mongolian ‘golden spring’) in Northern Mongolia. ‘Maimacheng, also Maimaicheng and other spellings, ‘Trading Town’ in Chinese or Naimaa Hot in Mongolian. Chinese trading towns with this generic name occurred at several points along the frontiers of the Manchu empire’ (Avery 2003: 178).

Chinese Pidgin Russian 3

was the only place where merchants from both sides conducted their business. !e "rst source that mentions ‘broken Russian’ (apparently CPR as it existed in the 18th century) is Pallas (1776: 134), who visited Kyakhta in 1772:

All the Nikans5 who do their business here on the border understand Mongolian,6 in which language the Russian merchants also o%en deal with them, either di-rectly or through a hired or sworn interpreter. Very many of them also speak broken Russian, but their pronunciation is so so% and faulty that a newcomer has di#culty understanding them.

!e next important source is Timkovskiy (1824: 66), who visited Kyakhta in 1821–1822:

It is proven by experience that the Manchus and the descendants of the Mongols who joined the Manchus in the conquest of China are much more skilful in pro-nouncing Russian words than the Chinese. !is can be proven by the corrupted and ridiculous dialect, in which Shanxi7 merchants make important transactions

5. German Nikaner, R nikancy < Manchu niqan ‘Chinese, boor’; Mongolian and Tungusic (Negidal, Oroch, Ulcha, Orok, Udihe) cognates also mean ‘servant, slave’ (Tsintsius et al. 1975–1977: 590, 637; Lewicki 1949: 31; Mudrak). !is term was used to denote the conquered Han Chinese as opposed to to the Manchus and Mongolians (Demidova & Miasnikov 1969: 497).

6. Mongolian was used as a language of o#cial and commercial communication between China and Russia in the 17th-19th centuries (Demidova & Miasnikov 1969: 518). Mongolian was taught at the Russian-Mongolian Schools in Irkutsk (provincial centre for Kyakhta) since 1725 (Yan 2006: 95), in Kyakhta since 1833, in Selenginsk since 1851 (Kyakhta 1990).

7. Shanxi (山西), a northern province that included parts of Inner Mongolia at the time and was famous for its merchants and bankers (晋商) (Lufrano 1997: 9). Not to be confused with the

!e area of Chinese Pidgin Russian.

!"#$%&''()

4 Roman Shapiro

with our merchants, who never learn Chinese. E.g. they call a horse [R loshad’] loshka; they say zames’ac for together [R vmeste];8 they say F’otel for F’odor [a male given name] and tuaceti p’ati monisa for twenty-$ve coins [R dvadcat’ p’at’ monet]. Chinese merchants even compiled large dictionaries in this incomprehensible dialect.

We observe the same phonological processes — consonant drop, epenthetic vow-els, substitution of l for r — as in the later CPR (see Section 3). Whereas these features quite straightforwardly re$ect Chinese phonology and therefore might have developed independently in Timkovskiy’s pidgin and in the later CPR, id-iosyncratic vocabulary correspondencies (loshaka present in Shprincyn archive, moniza and mes’aca present in Cherepanov 1853) testify to the contrary. It is worth noting that every time Timkovskiy mentions the language of Chinese merchants he calls it ‘Kyakhta Russian-Chinese dialect’ (1824: XII, 66, 183–184) rather than something like ‘broken Russian’.

A slightly more detailed description of CPR was given in ‘A letter from Kyakhta’ [Moskovskij Telegraf 1831: 5] in 1831. !e article is unsigned; however, it has been attributed to a famous Russian sinologist Rev. Ioakinf (Nikita Yakovlevich Bichurin) [Skachkov 1977: 109]. !e author claims that the ‘Russian dialect’ was unilaterally created by the Chinese:

!e Chinese made a law that every boy sent to do business in Kyakhta should learn Russian as soon as he arrives here. To that end, every shop has a small Chinese dictionary with Russian translations written in Chinese characters. However as the Chinese language has neither declension nor similar words for the tonic [=phonetic?] rendering of our syllables that have two or three consonant letters, our friends’ translation has formed a special Russian dialect, where our words, pronounced in a crippled way, are used as they are shown in the dictionary, with-out a hint at declension.

!e article also cites "%een examples of the ‘Russian dialect’ (see the Appendix). !ey consist of accurately reproduced Chinese syllables, which means that they must have been originally transcribed in Chinese characters and taken from a ‘dictionary’ (rather a phrasebook) of the type described in the Moskovskiy Telegraf 1831.9

neighbouring province of Shaanxi (陕西). In the Russian original Timkovskiy spells it Шааньси (Shaan’si) in Cyrillic, but Chansi in the Roman (French?) transcription.

8. Zames’ac sounds quite similar to the Russian expression za mes’ac ‘for a month’. !e Russian words for ‘month’ and ‘together’ may have originally merged in one CPR word in a neutralising context like ‘let’s settle our monthly accounts / all accounts taken together’.

9. Bichurin was one of the creators of Cyrillic transcription for Chinese, and its correct use is probably the main linguistic argument for his authorship of Moskovskiy Telegraf 1831. !e


Judging from the limited data available, the ‘Russian dialect’ must have been grammatically similar to later versions of CPR: it had no in$ection except the past/inceptive verbal forms that have double Russian/Chinese etymology (see below), and it placed the predicate in the "nal position and most of its verbs ended in i/j (the last feature is not really informative though, as this su#x corresponds with the Russian imperative form and many sentences are imperative here).

With development of trade CPR gradually spread along the whole Russian-Chinese border (valleys of the Shilka and Amur rivers, towns of Sretensk, Blagoveshchensk, Habarovsk.). A%er the Chinese Eastern Railway (CER or KVZhD) was constructed in 1897–1901 in the Russian Far East and Eastern China, CPR spread all over the railroad and in the city of Harbin, where it served as the main means of communication between the Russians and the Chinese, though both Russian and Chinese were taught as foreign languages (Oglezneva 2007a: 100–101). !e expansion across vast territories and among various indig-enous groups (the Koreans, Nanai, Udihe, Taz10) led to grammatical and lexical variation and the emergence of ethnic varieties. !e 20th century saw the slow evolution of CPR towards standard Russian due to the in$ux of Russian popu-lation to the Far East and universal schooling in Russian. Nowadays CPR rudi-ments can be found with the older generation of Siberian indigenous peoples and in mixed Chinese-Russian families where the Chinese did not acquire Russian. Russian-Chinese contacts in the Far East intensi"ed in the 1990s and gave rise to a (pre-)pidgin (brie$y discussed below). !is paper tries to focus on ‘CPR proper’ (i.e. Chinese varieties); however data of indigenous varieties are used sporadically to make up for the scantiness of ‘Chinese’ data.

2.1 Scholarship on CPR

!e "rst attempt at linguistic analysis of CPR was made by Cherepanov (1853). His article also included sample sentences and two texts, which are placed in the Appendix to this paper. Here are Cherepanov’s ideas on the origin of CPR:

1. !e Chinese who wished to trade with Russia were allowed by their govern-ment to do so only a%er passing an examination in ‘Russian’. !is claim is cor-roborated by the Moskovskiy Telegraf (1831) (see quotation above) and Maksimov (1871: 490–492; he made his jouney in 1855), who says that CPR was taught in Kalgan (Mongolia) and that the Chinese:

transcription is rendered in the paper through the Chinese Phonetic Alphabet (汉语拼音字母 han4yu3 pin1yin1 zi4mu3).

10. !e Taz formed in 1880s from intermarriages between the Udihe, Nanai and Chinese.

!"#$%&''()

6 Roman Shapiro

made a fundamental state law that a merchant should get the right to engage in trade only a%er he has learnt to read and write Russian. A secret instruction from Beijing says: ‘!is measure is necessary to prevent the Russians from learning Chinese, as if they know it they can "nd out the secrets of our trade and politics’. So we have not met a single Russian in Kyakhta, who could speak Chinese. And no Chinaman can get a passport and be allowed on this side of the wall until he can recite in a row and at random the long vocabulary brought from the Russian border, until he can write this with Chinese characters and has sat an exam in both these skills. !us all the Chinese there break the Russian language, and all Russian merchants try to "t in with this broken speech known by the name of Kyakhta trade language.

In 1852 the newspaper Moskovskie Vedomosti published secret instructions sent to the dzarguchey (the senior Chinese o#cial in Kyakhta) by the central gov-ernment. !e main idea is that Chinese merchants should coordinate their ac-tivities under the direction of the dzarguchey, act in the interests of the Chinese state and keep their maneuvers strictly secret from the Russians. !e last para-graph runs as follows: ‘It is forbidden for newcomers to Kyakhta to do business in person for a year, even if they speak Russian, lest they inadvertently disrupt the coordination of a&airs’ (quoted in Korsak 1857: 335).11 !is seems to imply that knowledge of ‘Russian’ was necessary (though not su#cient) to engage in trade in Kyakhta. !is is supported by the general attitude of the Chinese gov-ernment towards foreigners, particularly before a series of treaties were signed as a result of the Opium wars (i.e. before the 1840s-1860s). Cf. the 1721 and 1724 edicts banning missionaries and consequently all foreigners from China and mak-ing it a capital o&ense to teach Chinese to foreigners (Dan 1966: 22) or a regula-tion governing foreign trade under the ‘Canton System’ (1760–1842): ‘Foreigners may neither buy Chinese books, nor learn Chinese’ (Hsu 2000: 201). Russians still had some access to Chinese, even locally; a Russian-founded college of Chinese existed in Kyakhta since 1835. However, the sources indicate that this access was quite restricted. As for a Russian school in China, besides a school in Kalgan, men-tioned by Maksimov, we know that the Russian College (俄罗斯文馆) functioned in Beijing since 1708 (Masini 1993: 5, Yan 2006: 96).

2. According to Cherepanov, special textbooks and dictionaries were produced for examination needs since the early 18th century (he said in 1853: 371 that ‘the irregularities have not changed for the last 150 years’). !ey played an important role in the codi"cation of CPR. CPR was stable ‘unlike the idiom of other for-eigners, whose mistakes are accidental and change in the process of acquiring the [Russian] language. !e Russians also use this broken language when speaking

11. Korsak clari"es on the same page that the Chinese speak ‘corrupted Russian’.


with the Chinese’ (Cherepanov 1853: 371). !e fact that phrasebooks existed in the 19th century is supported by a number of sources ("rst mentioned by Timkoskiy, who visited Kyakhta in 1821). !ese sources also share a set of CPR features (pho-nology, vocabulary, SOV order), which supports the idea that CPR was quite stable in the 19th century. Phrasebooks must have contributed to this stability, since the Chinese community in Kyakhta was subject to migration shi%s:12 Chinese mer-chants were not allowed to bring their families with them; they went home a%er a few years in Kyakhta and young merchants were sent here from China rather then brought up locally (Moskovskij Telegraf 1831, Cherepanov 1853, Maksimov 1871: 492). We know that textbooks of Russian were produced in China in the "rst half of the 18th century (Yan 2006: 1996). What we do not know is whether Cherepanov saw these 18th century textbooks and based his ‘150-year stability’ claim upon comparison of their data with what he was hearing. On the one hand, his CPR description is generally reliable, as it shows a consistent pidgin system and is con"rmed by other sources, so he probably had some grounds to make this particular claim. On the other hand, we need more evidence to draw any conclu-sions about the existence and form of CPR in the 18th century.

Besides the Russian works quoted above, some CPR material can be extracted from papers by Alexandrov (1884) and Vrubel (1931). Shprintsyn collected pid-gin data in the Russian Far East (mainly the Primorye region) in 1929 and 1930 (Shprincyn 1932, 1968, archive). Perekhvalskaya and Belikov collected new materi-als in 1985, 1990, and 2004 and published a number of works (Perkhvalskaya 2003, 2004, 2008, Belikov 1997). Yang Jie published an article (2007) on some aspects of a Russian-Chinese ‘pre-pidgin’ that has been developing on the Russian-Chinese border since the late 1980s. Yang Jie’s data pertain to the variety spoken around the towns of Zabaykalsk (Russia) and Manzhouli (China). Oglezneva (2007a,b) describes the variety spoken in Blagoveshchensk (Russia) and Heihe (China).13 !e few examples show a typical CPR phonology (see Section 3) and elements of verbal morphology: in most cases the verb tense and number (but not person) are used correctly from the point of view of Russian, however the noun is not declined.14 !is kind of morphology makes the ‘pre-pidgin’ similar to the Taimyr Russian Pidgin, or ‘govorka’ (Xelimskij 1987, 1996, 2000, Stern 2002, 2005a).

12. For migrations of Russian merchants and their etholinguistic make-up see Stern 2005: 180, Korsak 1857.

13. E. Oglezneva’s book also contains information on the history of pidgin and Russian proper in the Far East, as well as a considerable appendix with fragments of travelogues and memoirs.

14. !is is true of the Chinese who speak the ‘pre-pidgin’. !e Russian speakers use Russian phonotactics and basic (but mostly correct) Russian grammar. !e pidgin features are CPR

!"#$%&''()

8 Roman Shapiro

Verbs usually take the "nal position. !ere are no sentence-"nal markers of tense/aspect/negation, which are typical of the older CPR. !e degree of genetic and typological relatedness between the older CPR and the pre-pidgin will be clearer when (and if) the latter stabilises.

vocabulary (including the famous CPR word kapitana ‘boss’) and a tendency to omit prepositions.

A dialogue in Chinese Pidgin Russian from Maksimov (1871: 488).


Western scholars who worked on CPR include Schuchardt (1884), who com-pares CPR with other pidgins; Jabłońska (1957), who gives a brief survey of CPR as spoken by the Chinese and Poles along the Chinese Eastern Railway in the "rst half of the 20th century; Neumann (1966), who concentrates on Cherepanov’s article; Nichols (1980, 1986), who studies pidgin fragments in Russian literature; and Stern (2005), who tries to shed light on the external history of Kyakhta pidgin formation.

Table 1. Pronunciation and spellingRussianpronunciation

Russianspelling15

Chinesetranscription

Meaning

1. /od’in/16 odin (один) eding one2. /fchira/ vchera (вчера) wuchila yesterday3. /!tava/ !togo (этого) edika this4. /prishol/ prishel (пришел) bulishela came5. /xadí/ xodi (ходи) huodi come, walk6. /prasí/ prosi (проси) buluoxi ask7. /tavár/ tovar (товар) duwal[e] commodity8. /xaróshij/ xoroshij (хороший) hulaoshi good9. /cyná/ cena (цена) caina price

Non-academic sources include works of literature depicting contacts between Russians and indigenous peoples of Siberia, who o%en spoke a variety of CPR (Arsenyev 1983, Maksimov 1871, Veresaev 1986 etc.) and textbooks of ‘Russian’ compiled for Chinese merchants going to Russia.

2.2 Spelling norms and lexicon

!e compilers of the ‘dictionary’ quoted in the Moskovskiy Telegraf 1831 (the earliest examples we have) obviously based their consonants on written Russian:

15. !e following latinization of Cyrillic is used throughout the article, except references, as well as Jabłońska and Schuchardt’s data, where their transcription is retained: a-a; б-b; в-v; г-g; д-d; е-e; ё-’o/jo; ж-zh; з-z; и-i; й-j; к-k; л-l; м-m; н-n; о-o; п-p; р-r; с-s; т-t; у-u; ф-f; х-x; ц-c; ч-ch; ш-sh; щ-q; ъ-`; ы-y; ь-‘; э-!; ю-‘u/ju; я-‘a/ja.

16. Russian dental stops are strongly palatalised before /i/ and (in most cases) /e/. In this posi-tion they resemble Chinese palatal a&ricates rather than dental stops, which are never palatal-ised. Cf. modern Chinese transcriptions of the former and current Russian presidents’ names: 普京 Pu3jing1 Putin and 梅德韦杰夫 Mei2de2wei2jie2fu1 Medvedev. Numbers here and else-where in Mandarin transcriptions stand for lexical tones.

!"#$%&''()

10 Roman Shapiro

there are a few transcriptions that render features of Russian spelling as distinct from pronunciation. As for the vowels, they are mostly quite di&erent both from Russian spelling and pronunciation. However, there are three examples (4–6 in the Table 1) that may re$ect the spelling. !e language of the dictionary is clearly pidginized; it is very much ‘wrong’ from the point of view of Russian grammar and choice of vocabulary.

Flug (1935) gives a short account of the book Russian foreign language (俄罗斯番语 E2luo2si1 fan1yu3). It includes a dictionary and a phrase-book, from which two sentences are quoted (Chinese Phonetic Alphabet used here, as for the Moskovskij Telegraf 1831):

(1) Du-lu-guai ma-nian-er nian-du. (CPR)17

Drugoj maner netu18

other manner NEG ‘!ere is no other kind [of goods]’.

(2) Di-bi ge-dao-er-li chi-lao-wei-ke? (CPR) Tebe kotoryj chelovek? you which person? ‘Where do you come from?’

[Compare the similar structure of: a. a Chinese set phrase 你 (是) 哪里人? Ni3 (shi4) na3li ren2 you (be) where person b. Arsenyev’s sentence: Tebe kakoj l’udi? you which person]

Shprintsyn (archive) gives a more detailed account of four textbooks. Russian and Chinese conversation 中俄话本 Zhong1 E2 hua4ben3 (1902) and Chinese and Russian conversation for beginners 华俄初语 Hua2 E2 chu1yu3 (6th edition 1927) tried to imitate standard written Russian and were only partially pidginized. !ey were typeset and their pages contained Russian words (in Chinese characers and Cyrillic) on the right and their Chinese translations on the le%. !e transcriptions of Zhong E (1902) were more precise and were probably based on the Russian spelling (cf. 歌耳五得 ge1 er3 wu3 de2, R[ussian] grud’ ‘chest’; 隈铁耳 wei1 tie3 er3,

17. Here and elsewhere we try to gloss Russian etymons rather than pidgin words per se. !ere are three reasons for this: (a) pidgin meanings are quite broad and vague; (b) they are usually clear from the context provided; (c) knowledge of more speci"c Russian meanings will be help-ful for non-Russian-speaking scholars should they wish to study CPR.

18. Russian etymons are given in this line.


R veter ‘wind’); Hua2 E2 was actually a revised version of the "rst book, but its transcriptions were more similar to the real pronunciation of Chinese who spoke Russian (cf. 歌鲁得 ge1 lu3 de2 ‘chest’; 曰结拉 yue1 jie2 la1 ‘wind’). !ese books were divided into chapters: ‘Military conversations’, ‘Conversations on railway contracts’, ‘Conversations of merchants’, etc. Appendices included the Russian al-phabet, as well as the tables of Chinese syllables and last names (as pronounced in the Shandong dialect) transcribed in Cyrillic.

"e complete Chinese and Russian conversations, revised and corrected (清俄会语俱全校正无讹 Qing1 E2 hui4yu3 ju4quan2 jiao4zheng4 wu2e2, Qing dynasty, date unknown) and Chinese and Russian vernacular conversations (中俄通俗话本 Zhong1 E2 tong1su2 hua4ben3, from 1913) were genuine sources on CPR. !ey were block-printed and contained only Chinese characters, which were used both for Chinese and Russian words/phrases. !e transcription was similar to how the Chinese actually pronounced Russian words: xiguluo for R steklo ‘glass’, ligasi for R lekarstvo ‘medicine’ (Shprincyn 1967: 90). !ematic grouping was less techni-cal than in bilingual books: Land and Buildings; Vegetation; Clothes; Body Parts; Industry and Trade; Commodities; Verbs; Various Words, etc. !ese textbooks in-cluded some literary Russian words, which were probably taken from the "rst class of editions. However the bulk of their vocabulary was colloquial and could be divided into the following etymological categories:

a. Common Russian: sidawaiya < R vstavaj ‘get up’; Duoya madamu yexi? ‘Are you married?’ (lit. ‘Your madam is?’)

An interesting subcategory is formed by Russian words that altered their mean-ing in CPR: wa3guan1na4 ‘train’ < R vagon ‘carriage’; ke4gu3shi2ke4 ‘clock’ < R kukushka ‘cuckoo’; wei2suo3ke4 ‘sun’ < R vysoko ‘high’; she2xi1bu4li4 ‘revolver’ < R shest’ pul’ ‘six bullets’ (a synonym: li4wo4li4ke4 < R revol’ver); a1la1xin1 ‘to close one’s mouth’ < R razinut’ ‘to open wide [one’s mouth]’ (sic!); zai4guan-1na4 ‘married’ < R zakonnaja ‘lawful [wife]’.

b. Siberian dialects: zha2yi2mu4ke4 < R zaimka ‘small settlement’; yi3zu2bu4li4 < R iz’ubr’ ‘Manchurian deer’; suo3bu4ke4 < R sopka ‘(volcano) hill’; yi2chi3ji2 < R ichigi ‘ladies’ boots’; wa3er3na4ke4 < R varnak ‘Siberian exile’.

c. Cossack dialects: a1di4man4na4 < R ataman ‘Cossack chie%ain’; ge4zha2ge4 < R kazak ‘Cossack’; si4da4ni2zi < R stanica ‘Cossack village’; ka3la1wa3 < R golova ‘chief ’.

d. Ukrainian borrowings: ni2jie2li4 < Ukrainian nedil’a ‘Sunday’; bao3guo4dan1 < Ukrainian pogoda ‘good weather’.

!"#$%&''()

12 Roman Shapiro

3. Phonology

!e analysis of CPR phonology is limited by transcriptions of the sources, which are prone to be adjusted either to Chinese or to Russian pronunciation and writ-ing systems. Shprintsyn (archive) paid more attention to the phonological side of CPR. !is paragraph is primarily based on his data and some of his ideas, though the main features described here (opening of syllables, merger of voiced and voice-less stops and a&ricates) are typical of other varieties as well.

Either Chinese or Russian phonological inventory was used depending on a speaker’s native language. Russians tended to pronounce pidgin words like their Russian etymons and use elements of Chinese phonotactics only sporadically (CPR liba < R ryba ‘"sh’, CPR dyva < R dva ‘two’). In the few words of Chinese origin Russians substitute /n/ for /ŋ/, /z/ for /dz/ (e.g. CPR fanza < C[hinese] 房子 fang2zi ‘house’), unaspirated voiceless consonants for aspirated consonants (CPR kanikan < C 看一看 kan4 yi1 kan4 ‘look’) and voiced or voiceless consonants for unaspirated consonants (CPR dadady/tatady < C 大大的 da4 da4 de ‘big’).19

!e Chinese had to make more changes to adapt originally Russian words to their native phonology. Two strategies are used to avoid consonant clusters, separately or simultaneously: (1) consonant drop, and (2) epenthetic vowels. Consonant drop is used as a rule in clusters of more than 2 consonants (CPR salaf < R shtraf ‘"ne, penalty’), in clusters [v (>w) +stop] (CPR lem’a < R vr’em’a ‘time’), in clusters [(dental) stop + n] (CPR sivoni < R sevodn’a ‘today’), in clusters [l + stop], with vocalisation of the Russian velarised [l] (CPR voŭk < R volk ‘wolf). Shprintsyn (archive) also mentions dropping /n/ in the syllable ni (CPR izynaj < R n’i znaj ‘don’t know’). !is feature may be compared with dropping palatalised n before i in certain contexts in some Chinese dialects (cf. 砚 Beijing yan 4, Chengdu ńian 3 ‘inkstone’). Epenthesis is used elsewhere to deal with consonant clusters. In non-"nal positions the epenthetic vowel tends to coincide with the stressed vowel (CPR seneka < R sneg ‘snow’, CPR sibichka < R spichka ‘match for producing $ame’) or accommodate to the preceding labial or velar consonant (CPR bulishola < R prish’ol ‘come’, CPR igylaj < R igraj ‘play’). At the absolute end of a word the epenthetic vowel used to open the syllable is generally -a a%er a non-palatalised consonant (CPR taka < R tak ‘so’) and -i a%er a palatalised consonant (CPR soli < R sol’ ‘salt’); -u is possible a%er labials (CPR padomu < R potom ‘later’). !e Russian postalveolar trill r is rendered as l (CPR pasmali < R posmotri ‘look’).

19. In the Chinese Phonetic Alphabet the characters that are normally reserved for voiceless stops and a&ricates stand for relevant aspirated sounds, while voiced letters render non-aspi-rated sounds.


!e Russian voiced and voiceless consonants merge into Chinese unaspirated half voiced consonants (CPR padomu < R potom ‘later’).

No tone has been found in CPR. Jabłońska (1957) suggested a tone in CPR xochu ‘want’; however, this idea was criticised by Shprincyn (1967). We would suggest that lexical tone is unlikely in contact languages with non-tonal languages as their lexi"ers (e.g. in CPR lexi"ed with Russian): what could be the principle of ascribing a particular tone to a given CPR word derived from a non-tonal and non-morphosyllabic Russian word? Another important factor is probably the presence or absence of tone in the mother tongue of contact language speakers. Compare the regular tonal re$ection of Chinese tone in Chinese loanwords (and morphemes) in Vietnamese and its loss in Chinese borrowings into Japanese — Chinese and Vietnamese being tonal morphosyllabic languages, unlike Japanese.20 Tones are found in Singlish (Lim 2009, Wee 2007), but only loans from Chinese seem to possess real lexical tones, whereas other phenomena may be better de-scribed in terms of pitch accent (as in Japanese, Serbian, or Lithuanian) and in-tonation. Perekhvalskaya claims that a Chinese-based pidgin of the Primorye indigenous population (Russian Far East) has two tones (2008: 32). However, her work on the subject remains unpublished.

!e CPR stress pattern usually follows that of Russian, though there is some over-generalization (i.e. CPR chetyrnácat’ < R chetýrnadcat’ ‘fourteen’ (Shprincyn archive), as Russian numerals 12, 13 and 15–19 all end in -ácat’ and have penulti-mate stress). Cherepanov says that ‘words are pronounced syllable by syllable and the stress is either absent or is placed on a middle syllable just for the sake of sim-plicity: molóko cf. Russian molokó, zharéko, cf. Russian zhárko, duráki cf. Russian durakí’ (Cherepanov 1853: 373). However, like most other CPR scholars, he omits the stress in his transcriptions.21

4. Morphology

CPR has some basic morphology, though neither declension nor conjugation exist. It is possible to draw a rough formal distinction between verbs and non-verbs (among

20. We are not making a very strong statement that Vietnamese and Japanese are contact lan-guages; however, an analogy is permissible here.

21. !e attentive reader may have noticed that sometimes CPR retains Russian phonological features that it should have lost according to the formulated rules: sibichka, pasmali, zhareko etc. However it should be remembered that the CPR data were recorded by a native Russian speaker, who was prone to ‘correct’ them. !erefore deviations from Russian are far more informative for us than transcriptions that are correct from the point of view of Russian.

!"#$%&''()

14 Roman Shapiro

notional, i.e. non-syntactic words). !is morphological distinction is supported by the semantic (verbs denote actions or states, while non-verbs dentote objects, quali-ties etc.) and syntactic criteria (verbs tend to take the last position in the sentence).

4.1 Non-verbs

Most words derived from Russian nouns (145 out of 193 in Perekhvalskaya’s vo-cabulary) end in -a:

kapitana ‘Western man, important person’, lisa ‘rice’, kapusa ‘cabbage’, salata ‘sal-ad’, liba ‘"sh’ (Shprintsyn archive); m’asa ‘meat’, leba ‘bread’, chena ‘money, price’, syna ‘son’, baryshn’a ‘miss’, leta ‘summer’, dlova ‘"rewood’, sonca ‘sun, day’ (Vrubel 1931); madama ‘Western woman’, kunja ‘daughter, girl’,22 liesa ‘forest’(Jabłońska 1957); kilometla ‘kilometer’, chasa ‘hour’, masina ‘car’, lekasta ‘medicine’, magazina ‘shop’(Kukan).

!e -a su#x has several sources. !e vast majority of Russian nouns in their de-fault form (usually nominative singular) end either in a consonant or in -a, -o, -e. All the three vowels, when unstressed, are reduced into a kind of schwa. !e Russian nouns that end in a consonant usually add an -a in CPR to open the "nal syllable. !is process is also supported by the fact that these nouns have a Russian genitive singular (and in case of animate nouns also an accusative singular) in -a (Schuchardt 1884). Some Russian nouns that have nominative plural as their default form (as they do not have a singular or are more widely used in plural for semantic reasons) also end in -a. !e -a su#x in CPR sources must be unstressed in most cases, judging by Cherepanov 1853, Shprintsyn 1932, 1968, archive and Jabłońska 1957, where stress is shown. Perehvalskaya also argues that ‘the nouns of the Siberian pidgin are characterised by quite stable stress that is usually penul-timate and is only in rare cases placed on a last syllable (drova ‘"rewood’) or a pre-penultimate syllable (dozhika ‘rain’)’. An indirect argument supporting this point is also the phonology of Russian, as CPR normally retains the Russian place of stress (Shprincyn archive). In Russian, stress obviously falls on the stem in nouns with the zero nominative singular ending (i.e. nouns that end in a consonant), thus the epenthetic "nal -a in correspondent CPR words must be unstressed. All the above makes one think that what is transcribed with -a in CPR sources is most probably pronounced as schwa. !en the CPR words that originate from Russian adjec-tives or adverbs can be included in the same formal class of ‘non-verbs’: their "nal vowel is transcribed as -a or -!, which stand for schwa-type sounds in unstressed

22. !is word actually comes from Chinese (姑娘gu1niang ‘girl’) and is altered so that it follows the general rule.


positions in the Russian spelling system (actually it does not have a special letter for schwa). Here are some examples of di&erent etymologies of the CPR -a su#x:

– manera (Cherepanov 1853) ‘manner’< R manera (the nominative singular -a ending present in Russian);

– vorota (Cherepanov 1853) ‘gate’< R vorota (the nominative plural -a ending present in Russian; this is a pluralis tantum noun in Russian); correct

– rukava (Cherepanov 1853) ‘sleeves’< R rukava (the nominative plural -a end-ing present in Russian; this noun has a singular in Russian, however the plural form is in a wider use for semantic reasons)

– muzha (Cherepanov 1853) ‘husband’ < R muzh (the epenthetic -a added to a Russian consonant-"nal word, supported by the Russian genitive/accusative singular form in -a)

– sonca/shonsa/sólenca (Schuchardt 1884, Vrubel 1931, Shprintsyn 1968, Kukan, Jabłońska 1957) ‘sun, day’ < R solnce (the Russian unstressed -e reduced into a schwa)

– solovo (Cherepanov 1853) ‘word’ < R slovo (the Russian unstressed -o reduced into a schwa)

– neperemena (Cherepanov 1853) ‘certain(ly)’ < R nepremenno (the Russian un-stressed -o in an adjective/adverb reduced into a schwa)

Other examples of CPR non-verbs that come from Russian adverbs and end in -a /schwa in CPR include: des’a (Kukan) ‘here’ < R zdes’; doma (Shprintsyn archive) ‘at home’ < R doma; niza (Shprintsyn archive) ‘below’ < R niz; šýpka ‘very’ (Jabłońska 1957) < R shibko; skoko (Kukan) ‘how many’ < R skol’ko; toko (Kukan) ‘only’ < R tol’ko.

34 nouns out of 193 in Perekhalskaya’s list end in -za/-ca, which is a re$ection of a dental "nal consonant in the Russian etymon, and at the same time this ele-ment resembles the Chinese noun su#x -zi (子):

jajca (Shprintsyn 1968) ‘egg’ < R jajco; (142) kitajca (Kukan) ‘Chinese’ < R kitajec; (143) lipaxoza (Kukan) ‘logging enterprise’ < R lespromxoz; (144) moneza (Cherepanov 1853) ‘money, coin’ < R moneta; (145) ledeneza (Cherepanov 1853) ‘lollypop’ < R ledenec.

37 nouns out of 193 end in -ka. -k- is a widespread Russian noun su#x (though surely not the only one); however, in CPR it may also be appended in cases when -k- is absent from the Russian etymons:

batinka (Vrubel 1931) ‘shoe(s)’ < R botinok; kapeka (Vrubel 1931) ‘kopeck’ (a Russian coin) < R kopejka; pampushka (Kukan) ‘Chinese steamed bun, baozi’ < R pampushka; rabochka (Shprintsyn archive) < R rabochij ‘worker’; loshaka (Shprintsyn archive) ‘horse’ < R loshad’.

!"#$%&''()

16 Roman Shapiro

Nouns that are derived from Russian plural forms (21 in Perekhvalskaya’s vocabu-lary) end in -i/-y (a%er unpalatalised/palatalised consonants): manty ‘dumplings’ < R manty < C 馒头 man2tou; l’udi (Arsenyev 1983, Kukan) ‘people, person’ < R l’udi; (153) panty (Kukan) ‘horns of a young deer’ < R panty. !is ending may have actually been a schwa rendered as -i/-y due to the in$uence of Russian spelling.

4.1.1 PronounsCPR personal pronouns form a semantic subclass of non-verbs.

Table 2. CPR PronounsSg etymology Pl etymology

1 maja/ja R possessive 1SG feminine mojaR personal 1SG ja

nasha R possessive 1 Pl feminine nasha

2 tibi/tvaja R personal 2SG indirect stem teb-R possessive 2SG feminine tvoja

vasha R possessive 2 Pl feminine vasha

3 ivo R personal 3SG genitive/accu-sative or possessive 3SG ego

ivo R personal 3SG genitive/accu-sative or possessive 3SG ego

1. !e 1sg form maja is the only form used in Cherepanov 1853, Aleksandrov 1884, Schuchardt 1884, Jabłońska 1957, Shpritntsyn, Arsenyev 1983. !e form ja is present only at the end of the 20th century (Kukan). Maja is argued to have a double etymology from R 1SG poss fem moja and from English my. It presumably originates in the Russian-English pidgin of the 17th century de-scribed in Alekseev 1968. Perekhvalskaya bases her argument on a single sen-tence Gimi moja foro scrivers < English give me + moja + English for scriveners. She explains moja in this sentence as a personal 1SG pronoun that repeats the English me of gimi (Perekhvalskaya 2008: 127). However, it is also possible to interpret it as a possessive 1SG pronoun. In this case the sentence translates as ‘give me my quill’ rather than ‘give me a quill’.

2. !e forms tvaja, vasha, nasha, and ivo, which originate from Russian pos-sessive pronouns (ivo can also be connected with the personal 3sg genitive/accusative form), are explained as a generalisation of maja (Perekhvalskaya 2008: 128). Schuchardt (1884: 320) thought that the use of possessive pro-nouns instead of personal pronouns is typical of pidgins with a Chinese com-ponent. He gives two examples: (i) Chinese Pidgin English my that functions as the English I, my, and me; (ii) ‘Chinese-Spanish’ (Philippines) mia quiele ‘I want’ instead of Spanish yo quiero. As for the use of the feminine instead of


the masculine, Schuchardt considers it a widespread phenomenon in pidgins in general and gives the following examples: (a) ‘Chinese-Portuguese’ (Macao) and ‘Indian-Portuguese’ [Goa?] minha, tua, sua instead of meu, teu, seu; (b) ‘Negro-Portuguese’ (Cape Verde) nha (<minha) caballo ‘my horse’ instead of Portuguese meu cavallo.

Chinese possessives have the structure ‘personal pronoun + de (的)’, so the con-nection between possessive and personal pronouns is quite transparent. Besides, in certain common contexts de tends to be dropped:

(3) 我的书 wo3 de shu1 I POSS book ‘my book’

(4) 我爹; 我家 wo3 die1 wo3 jia1 I Dad I home ‘my Dad’; ‘my home’

In Tungusic languages (Manchu, Negidal, Evenki), which could have in$u-enced CPR, possessives are also constructed regularly from personal pronouns (Möllendorf 1892: 6; Tsintsius 1982: 22; Bulatova 1999: 26). !is structure is also sometimes calqued in CPR with Chinese morphemes:

(5) Ni dy šýma múr-mur? (Jabłońska 1957: 167) you POSS what say ‘What are you saying?’

As for the female forms maja, tvaja, they must have been easier for Chinese speak-ers, because they can be approximated with Mandarin syllables ma-ya, te-wa-ya (Mandarin does not have the combination *-oi, which would be needed for ren-dering the Russian masculines moj, tvoj).

All these factors must have contributed to the development of forms maja etc. in CPR. Although there are enough explanations of these forms within the languages of the region, a parallel connection with a Russian-English pidgin form is quite probable, as moja is also present in Russenorsk and in the Russian set expression that parodies foreigners (not necessarily Chinese) speaking Russian:23

23. For a theory of Russian pidgin monogenesis, which is beyond the scope of the present paper, see Kozinsky 1974.

!"#$%&''()

18 Roman Shapiro

(6) Moja tvoja ne ponimaj. 1SG.POSS.FEM 2SG.POSS.FEM not understand.IMV ‘I don’t understand you’.

(7) Compare standard Russian: Ja teb’a ne ponimaju. 1SG.NOM 2SG.ACC not understand.1SG

It is also worth noting that pronouns, as any non-verbs, can be used both for theta roles and as modi"ers, i.e. as personal and possessive pronouns:

(8) Agent/topic: Za vashe zh’onusheki mes’aca posidi esa? TOPIC your wife together sit REL ‘Do you sit together with your wives?’ (Cherepanov 1853)

(9) Modi"er: Za vasha muzha zhesetoki manera pozhivu esa. TOPIC your husband cruel manner live REL ‘Your husbands live in a cruel manner [= are rude]’. (Cherepanov 1853)

4.2 Verbs

Most verbs are taken from Russian in the imperative singular form. Most Russian imperatives use the su#x -i or -j, so -i/-j has become the verbal marker in CPR.24 !e choice of Russian imperative as a default verbal form in CPR can be explained by a few factors. First, this is one of the most frequent forms in communication, especially between traders or employers and employees. Second, the Russian imperative o%en looks similar to the present/past tense verbal stem, cf. R govori ‘talk!’, govori[sh’] ‘you talk’, govori[t] ‘he talks’, govori[l] ‘he talked’, govori[t’] ‘to talk’, etc., therefore it was convenient to generalize. !ird, the fact that in Mongolian and Manchu imperative 2SG is the base for all other forms (Todaeva 1951: 98, Möllendorf 1892: 9) may also have played a certain role. Below are some examples of CPR verbs:

Podumaj ‘think’, poxodi ‘come’, pomeshaj ‘interfere’, posidi ‘sit’(Cherepanov 1853); kupí ‘buy’ (Schuchardt 1884), lamaj ‘break’, gavari ‘say’, pusykaj ‘put’, pomilaj ‘die’, zhivi ‘live’ (Shprintsyn archive), kushaj ‘eat’, nalej ‘pour’, bali ‘be ill’, sobiraj ‘collect’ (Kukan), rugaj ‘scold’, pamagaj ‘help’, iglaj ‘play’ (Vrubel 1931).

Most CPR forms reproduce Russian imperatives correctly, which means that they were borrowed as a whole; the Russian rules of adding the imperative su#x are

24. -i is used a%er a consonant and -j a%er a vowel.


quite complicated. However, there do exist some forms that were created within CPR, as they are di&erent from any Russian form:

poguli ‘have a walk’ cf. R imperative pogul’aj; potorgovaj ‘trade’ cf. R potorguj; xychi ‘want’ [no standard imperative in R] (Cherepanov 1853); iskaj ‘search’ cf. R iqi (Vrubel 1931); kushi ‘eat’ cf. R kushaj; sadi ‘plant’ cf. R sazhaj or posadi (Kukan).

!e few verbs of non-Russian origin also use this pattern (juli ‘row, work as a boatman’ (Vrubel 1931) < C 游乐 you2le4 ‘amuse’ [?]; kan’trami ‘kill’ < C 砍头儿吗 kan3 tou2r ma ‘cut o& one’s head’ — see under Vocabulary). Another widespread verbal su#x is -la (see Section 4.3 below). Other Russian verbal forms must have been borrowed later, at the post-pidgin stage. !ey are not used in early sources (Cherepanov 1853, Shprintsyn 1932, 1967, archive, Vrubel 1931).

!ere are only "ve exceptions to this rule in Cherepanov’s CPR: umesha ‘be able’ < R present 2nd person singular umeesh’; pozhivu ‘live’ < R future 1st person singular; serediza ‘be angry’ < R imperative singular serdis’ or in"nitive serdit’s’a with the re$exive su#x s’a; pixu ‘drink’, kushaxu ‘eat’ with a Mongolian element -xu added to Russian roots.25

While 128 verbs out of 166 in Perekhvalskaya’s (2008) vocabulary end in -i, modal verbs form a special subgroup; they do not end in -i and can be negated only by ne/ni, while other verbs can also take netu as their negation. !e following modal verbs can be singled out: nada < R nado ‘must’; mogu, mozheno < R mogu, mozhno ‘can, may’, ‘possible’; xochi, xychi < R xochu ‘want’, as in (16–19) below:

(10) Sama podumaj mozheno. oneself think possible ‘One could think about oneself ’. (Cherepanov 1853)

(11) Usytupi magu maja kupi. concede can my buy. ‘If you make me a discount, I’ll buy it’. (Shprintsyn archive)

(12) Skazyvaj ne nado. tell NEG need ‘No need to tell him’. (Cherepanov 1853)

25. According to Cherepanov, -xu (< Mongolian future participle su#x -h[u]) is completely lexi-"ed in CPR, it is used only in three words (the third one being ‘write’, once recorded as bichixu and once as bichi, from Mongolian) as a verb marker instead of -i/-j. Cherepanov is the only source for these words. Mongolian future participles in -h[u] are o%en used for representing verbs in dictionaries, which probably means that they are perceived as semantically most neutral.

!"#$%&''()

20 Roman Shapiro

(13) A govori ne mozheno, pishi adinak!. and speak NEG possible, write same. ‘[!e Chinese] can’t speak Japanese, but they write in a similar way’ (Kukan).

4.3 Sentence-"nal particles26

CPR has a set of particles, all of which take the "nal position in the sentence. Most of them stem from Russian verbs (there are also rarer variants that come from Chinese verbs). From the structural point of view their source is most probably Mongolian, where various tense forms of verbs ‘be’ and ‘exist’ are used as tense/aspect markers in sentence-"nal position (Todaeva 1951: 131), as in (14–15):

(14) Ter güjž bajna. he run be.NONPAST ‘He is running’.

(15) Sühbaatar 1923 ony 2-dugaar saryn 22-yn ödör nas barsan bilee Sühbaatar 1923 year second month 22-GEN day age deceased be.PAST ‘Sühbaatar died on February 22, 1923’.

!e fact that particles are sometimes used in sentences without verbs (examples 24, 27, 33) suggests that they modify the whole phrase/sentence rather than just the verb, much like the Chinese sentence-"nal particles (Chao 1968: 149). Moreover, one of the CPR particles is not overtly connected with a verb: la (expressing change of state) has its source in a Chinese particle and a Russian su#x. !e CPR particles are listed below with their rough meanings.

4.3.1 Tense/Aspect1. Budu/budi < R budu ‘I will be’ shows that the action will happen in the future (FUT). It is used in a few CPR sources:

(16) Posel’a zhalej budu. later regret FUT ‘You are going to be sorry later’. (Cherepanov 1853)

(17) Moja hodi budi my walk FUT ‘I will come’. (Schuchardt 1884)

26. Additional examples for the sections Sentence-$nal particles and Syntax are to be found under corresponding headings in the Appendix.


(18) Vasha des’a zhivi budu. your here live FUT ‘You are going to live here’. (Shprintsyn archive)

(19) Vada m!noga budi. water much FUT ‘!ere will be much water’. (Kukan)

2. !e ‘relevance’ [REL] particle esa, esi < R est’ ‘he is’, ju < C you3 有 ‘have’. !is particle is used in sentences describing (A) progressive activities, (B) habitual ac-tivities, or (C) results of activities — in this case esa/ju is o%en used together with the particle la. !e intersection of all these usages is probably the close connection with the time of speaking: (a) something is happening now; (b) something usually happens, and now it is happening, too; (c) something has happened before, and the result is important now. Note that the Chinese "nal particle 呢 ne has a similar range of meanings. Its function has been described as follows: ‘by using ne, the speaker draws the hearer’s attention to the information marked by the particle and urges the hearer to adjust shared common ground accordingly with regard to the current interaction’ (Wu 2005: 47). Here are three Mandarin examples (taken from Li Boya 2006: 7–21) for each of the above mentioned usages:

(20) A: progressive activities 说着话呢。 shuo1 zhe hua4 ne speak PROG word REL ‘!ey are talking — line busy’ [so I can’t make a call].

(21) B: habitual activities 他的英语说得比你还好呢。 ta1 de ying1 yu3 shuo1 de bi3 ni3 hai2 hao3 ne he POSS English language speak COMPL than you still good REL ‘He speaks better English than you’ [so don’t think he won’t get around in

the US].

(22) C: results of activities 香港最近下雪了呢。 Xiang1gang3 zui4jin4 xia4 xue3 le ne Hong Kong recently fall snow INCP REL ‘It snowed in Hong Kong lately’ [the listener wouldn’t have expected it].

Compare also a Mandarin usage, very similar to the CPR example (49):

!"#$%&''()

22 Roman Shapiro

(23) 今年春天呢, 我有幸. jin1 nian2 chun1tian1 ne wo3 you3 xing4 this year spring REL I have luck ‘As for this spring, I’m lucky’. (http://www.cnr.cn/nmgfw/xwzx/xwst/200805/t20080501_504776796_3.html)

For additional examples see Lü Shuxiang (2007: 412–413). Here are some CPR examples on the usage of this particle:

(24) A: progressive activities a. Poguli esa, gospodine Dalaj. walk REL Mr Dalaj ‘You are having a walk, Mr Dalai?’ (Cherepanov 1853) b. Nídy mašýynka ju. your (< C 你的ni3de) deceit REL ‘You are cheating’. (Jabłońska 1957)

(25) B: habitual activities Maja mala-mala chitaj esi. my little-little read REL ‘I read a little’. (Vrubel 1931)

(26) C: results of activities a. Synka pomilaj esi. son die REL. ‘[My] son has died’. (Shprintsyn archive) b. Jevo lamajla ju? he break.INCP REL ‘Has he fallen ill?’ (Jabłońska 1957)

It has been suggested that this particle may denote evidentiality. According to Nichols, ‘depending on tense-aspect and person, an evidential predication is ei-ther an inference from given evidence or an immediate reaction to it’ (Nichols 1986: 253). However, Aikhenvald (2004) has shown that though evidentiality can be marked simultaneously with tense-aspect, as well as with modality or mirativity, these categories are marked separately in some languages and therefore should be distinguished. !e following understanding of evidentiality may be more transpar-ent: languages with evidentiality mark whether the source of information is direct or indirect knowledge (some languages have subcategories for indirect knowledge, e.g. inferential, nonvisual, hearsay, etc.). From this perspective, immediacy, which suggests direct perception, and inference, should be opposed rather than consid-ered ‘contextual variants of a single category’ (Nichols 1986: 255). Nichols’ insight is very important; inference and immediacy do have something in common, but


this is probably relevance for the current interaction, i.e. direct connection with the present situation rather than evidentiality. Nichols states explicitly that her theory speci"cally describes the Ussuri dialect of CPR (Nichols 1986: 240), and indeed, the following example from (Cherepanov 1853) cannot be explained by evidentiality (in Nichols’ interpretation of the term):

(27) Za evo malen’ki kurema pode-bol’shana kurema ponosi esa. TOPIC he small jacket under-big jacket wear AUX ‘!ey wear a small jacket under a big jacket’.

Here a Chinese merchant in Kyakhta is describing in a letter what he habitually observes: this is what foreign gentlemen usually wear; neither inference from indi-rect sources nor immediate reaction to some new perception is present. However esa can be explained as a relevance marker here: ‘this is what they wear, and look now how strange they are!’ At the same time the examples used in (Nichols 1986) seem to "t in the ‘relevance’ framework.

3. !e inceptive [INCP] and past [PAST] particle la has a double etymology (both from R past tense27 su#x -l(a) and C aspectual su#x/particle le).

3.a. Inceptive usage!e term inceptive is used as a short approximation here, a more accurate inter-pretation is change of state. !is usage is much more ‘Chinese’ than ‘Russian’, both for semantic reasons and because the particle is appended to whole phrases rather than only to verbs.

!e usage in Mandarin can be explained by the following excerpt from a dia-logue (http://news.sohu.com/20050803/n226545546.shtml):28

(28) A: 终于又有希望了，我说我这几 Zhong1yu2 you4 you3 xi1wang4 le wo3 shuo1 wo3 zhe4 ji3 "nally again have hope INCP I say I this several

年总算没白费。 nian2 zong3suan4 mei2 bai2fei4 year actually NEG waste ‘And "nally there was some hope again [hope was ‘found’], and I said

that I hadn’t just wasted all those years’.

27. !e -l(a) su#x is used for all kinds of past tense in Russian. All aspectual distinctions (pri-marily perfective and imperfective) are rendered by pre"xes and other su#xes.

28. For more examples and discussion see e.g. Lü Shuxiang 2007, Li & !ompson 1981, Chao 1968, Li Boya 2006.

!"#$%&''()

24 Roman Shapiro

B: 这已经是哪一年了? Zhe4 yi3jing1 shi4 na3 yi1 nian2 le this already is which one year INCP ‘Which year was it [the year that marks the change of state: from ‘no

hope’ to ‘hope’]?’

!e INCP marking change of state in connection with years can be also found in the following CPR example:

(29) Kakoj gadu-la[,] a ni zn!j, ise ukalainza s’uda prishola. which year INCP I NEG know all Ukrainian here come ‘Which year was it, I don’t know when the Ukrainians came here’. (Kukan)

Compare also the following example (Taz):

(30) Tama sasema netu-la, s’o suda, trica doma. Staryj uzhe netu-la. there completely NEG-INCP all here, thirty house old already NEG-INCP ‘!ere are no more [Taz people] le% there, they have all moved here, thirty

families. !ere are no old [people] anymore’.

-la is also used to show the sequence of actions (a ‘pluperfective’ usage), as in Chinese (Kukan):

(31) Ivo l’udi zaxarani kucha-la ivo ni sadi, a qasa is’o takoj delaj. he person bury heap-INCP he NEG plant and now all such do ‘A%er they had buried someone under a mound, they did not plant anything

there, and now they do it’.

3.b. Past tense usageIn this case la has to immediately follow the verb. !is usage has parallels both in Russian, where -l(a) is a verbal su#x of the past tense, and in Chinese, where le can be used as a perfective verbal su#x.29 Semantically CPR is closer to Russian in this case, as the meaning is just past, not necessarily perfective. If we consider form, however, the CPR particle -la is usually appended to verbs in a ‘Chinese’ way, that is without any fusional processes characteristic of Russian: davaj-la ‘gave’, lomaj-la ‘broke’, zhivi-la ‘lived’, beri-la ‘took’ (Shprintsyn archive), pomiraj-la/poumilaj-la ‘died’, zaboli-la ‘got ill’, pish!-la ‘wrote’ (Kukan). In these cases -la is directly added to a Russian imperative form (which is the default verbal form in CPR) or to a present form (the last example).

On the other hand, some forms look exactly like their Russian counterparts: kupila ‘bought’ (Shprimtsyn archive), lechila ‘treated [an illness]’, uvozila ‘took

29. !e le1 (verbal, perfective) and le2 (phrasal, change of state) are distinguished in most stud-ies of Mandarin grammar. Chao (1968: 246) even gives them di&erent etymologies.


away’ (Kukan). !e above forms are also easily explained as a combination of a Russian imperative and -la. However, this is not always the case. Such irregular forms, e.g. nashol/nashola/nashla ‘found’, pashol ‘went’, ushla ‘le%’, pomil!/pomirala ‘died’ (cf. the regular CPR derivation pomiraj-la above), vypil! ‘drank’, byla ‘was’ (Kukan), and pulishola ‘came’ (Shprintsyn archive) must have been borrowed from Russian as ready words with -l(a). Here is an example of the past tense usage:

(32) !ta labodala dilektala. this worked director. ‘He worked as a director’. (Kukan)

In the following example the -la particle is added to a form that already contains -la as its constituent part, to imply a change of state:

(33) Uot !ta ush konchila-la, qasa netu. so this already ended-INCP, now NEG ‘We’ve already run out of it [ginseng], there’s nothing le%’ (Kukan).

CPR is not the only example of a language that has borrowed the -le particle from Chinese. !e particle is also widely used as a past marker in the southern Udihe dialects (primarily the Iman dialect): siki!mi ‘I washed’, siki!i ‘you washed’, siki!ni ‘he washed’ are all substituted by siki-l!. !e Udihe speakers perceive this feature as a Chinese rather than Russian borrowing (Perekhvalskaya 2008: 163).

4. !e particle byla/jula < R byl ‘he was’/ C 有(了) you3 (le) ‘had’. !e form and usage may also have been prompted by the Mongolian past auxiliary bilee ‘was’, cf. (66). It is not used very o%en. It seems to denote an activity or state in the past (PAST) that has no direct connection with the time of speaking and is basically synonymous to the more common -la is its past tense usage.

(34) Za vasha soledaza za nasha voseqanaka ponosi bylo, kako TOPIC your soldier TOPIC our cloth wear PAST how

mozheno bylo za evo pokoloti. possible PAST TOPIC he beat ‘Your soldiers were wearing our cloth, how could it be that they were beaten/

defeated’. (Cherepanov 1853)

(35) Karaula sypila jula, maja fanguli akyno. gards sleep.INCP PAST, my break window. ‘!e guards fell asleep, I broke the window’. (Shprintsyn archive)

!"#$%&''()

26 Roman Shapiro

4.3.2 Negation1. netu < R net(u) ‘no’, ‘there is not’: general negation (A), also used to state non-existence (B). While sentence-"nal negation is alien to Russian and Chinese it is common in Mongolian and Tungusic languages (Manchu, Nanai, Ulcha, Evenki).

(36) A: general negation !ta liba pomilaj netu. this "sh die NEG ‘!is "sh did not die’. (Shprintsyn archive)

(37) B: non-existence Synka netu. son NEG ‘[I] don’t have a son’. (Shprintsyn archive)

2. ni < R ne ‘not’: used before modal verbs. It is also a rarer alternative to netu with general verbs. !is negation is not "nal, it always precedes the negated verb, actually it should be considered an adverb.

(38) Adin ludi delaj kushi ne mogu one person do eat NEG can ‘One person can’t provide [for a family]’. (Vrubel 1931)

5. Syntax

5.1 Basic word order

In CPR, the only word classes distinguished in any formal way are verbs and non-verbs, and it would be most accurate to describe the word order as N1N2V, where N1 is usually the agent or topic, and N2 is any other type of thematic role. N2 is not necessarily the object, which is why N1N2V is preferred to the more conventional SOV formula here. N2 can also stand for a few di&erent semantic roles (cf. Jabłońska 1957):

(39) Jevó dúmaj majá jevó céna daváj. he (N1) think (V) I (N1) [he money] (N2, bene"ciary and patient) give (V) ‘He thinks I’ll give him money’.

However; this is rare; usually a sentence includes only two roles:

(40) X!dz’u liba kupila. !ta voda pusykajla Hejiu "sh buy.PAST this water let go.PAST ‘Hejiu bought a "sh. [He] let it into the water’. (Shprintsyn archive)


!e agent ‘Hejiu’ is present in the "rst sentence, but it is omitted in the second one, as it already has two roles, the patient and the direction. See the Appendix for other examples of using several sentences to show all the necessary roles of one verb, so that there are no more than two roles in each sentence.

!e CPR word order is quite di&erent from Russian and Chinese, which are both basically SVO. It may have emerged under the in$uence of Mongolian and Tungusic languages, which are SOV. Word order in NPs is also head-"nal both in CPR and in Mongolian/Tungusic (see Section 5.3).

5.2 !ematic roles and syntactic functions

All thematic roles can be denoted in CPR by nouns or noun phrases without any special markers. In this case thematic roles are simply deduced from the context, and there is no need to draw a distinction between thematic roles and syntactic functions. Word order plays only a marginal role here, as there is no formal way of distinguishing between e.g. Agent-Patient and Patient-Location.30

(41) Agent: Ivo p’ati goda bolshi! s’uda dzhivi. he "ve year more here live. ‘He has lived more than "ve years here’. (Shprintsyn archive)

(42) Patient: !ta zhen’shen’ kushaj xorosho. this ginseng eat good. ‘It is good to take ginseng’. (Kukan)

(43) Bene"ciary: Maja tvaja gavalila. my your said ‘I have told you [about it]’. (Vrubel 1931)

(44) Location: Des’a bal’niza byla, pomirala. here hospital PAST died ‘[She] died at the hospital here’. (Kukan)

30. As is well known, syntactic functions are coded with word order and syntactic words in Chinese and with cases, syntactic words and word order in Russian. For additional examples for this section see Perekhvalskaya 2008.

!"#$%&''()

28 Roman Shapiro

(45) Goal: Sólenca niétula, dózika ju, soledátka liésa pasóla. sun NEG.INCP rain have soldier forest went ‘!ere is no sun any more, it is raining, the soldier went to the forest’.

(Jabłońska 1957)

(46) Bul’dozera rovno delaj. bulldozer $at do ‘It was levelled with a bulldozer’. (Kukan)

(47) Transport: Potom jivo lodka hodi. then he boat walk ‘!en he le% by boat’. (Bikin)

(48) Source: A[ ]31 qasa toka kartoshka sadi, s’o magazine kupi a[ ] ran’she ne and now only potatoes plant all shop buy and before NEG ‘And they grow only potatoes now, they buy everything in the shop, it wasn’t

like this before’. (Taz)

(49) Time: !ta vesna est’, moja zapr’agaj budu. this spring REL my harness FUT ‘!is spring I will harness [my horse]’. (Shprintsyn archive)

Some varieties of CPR have speci"c role markers, as illustrated below.

5.2.1 Agent/topic markerCherepanov’s CPR has a speci"c agent/topic marker: za, as shown in the following:

(50) Ilisandera, za tibi kaka podumaj, za moja prish’ol esa? Alexandra TOPIC you how think TOPIC my come.INCP REL ‘Alexandra, why do you think I have come here?’

(51) Pocheto za vashi zhenusheki s’uda poxodi nitu? why TOPIC your wife here come NEG ‘Why don’t your wives come here?’

!ere are 26 examples of za marking an agent in Cherepanov (1853). !ere are only two cases (both within the same sentence) where za does not seem to mark a prototypical agent, however it can be considered a topic marker:

31. A space added by us.


(52) Za vasha soledaza[,] za nasha voseqanaka ponosi bylo, kako TOPIC your soldier TOPIC our cloth wear PAST how

mozheno bylo za evo pokoloti. possible past TOPIC he beat ‘Your soldiers were wearing our cloth, how could it be that they were beaten/

defeated’.

!is structure could be a calque of Chinese, with topics taking the "rst position in each short sentence:

(53) 你们的兵啊，我们的料子穿上了，怎么能 ni3mende bing1 a wo3mende liao4zichuan1shang4 le zen3me your soldier AUX our cloth put.on PAST

啊，他们打败了! neng2 PRT ta1men da3bai4 le how possible they defeat PAST!

(54) Tolika za direcheki odina beleshiki kusoka perevezhi esa. only TOPIC hole one white piece tie_up REL ‘Only in the hole, they tie up a white piece [of cloth]’.

Compare Chinese:

(55) 只有小洞 (里)，才系上一块白布。 Zhi3you3 xiao3 dong4 (li3) cai2 ji4shang4 yi1 kuai4 bai2 bu4. ‘Only small hole (in), only tie.up one piece white cloth’.

!ere are only three sentences in the dialogue where the secondary agent/topic is not marked by za. Two of them are dependent clauses introduced by kako (< R kak ‘how’). !e third one is a contraposition of two topics. In all three cases the main topics are marked by za:

(56) Za moja podumej, kako vasha zakona horoshaneki: zhenusheki TOPIC my think how your law good wife

mes’aza pozhivu esa. together live REL ‘I think your custom is so good: wives live together [with their husbands]’.

(57) Za moja Mikita skazyvaj budu, kako Dalaj pogovori esa. TOPIC my Mikita tell FUT how Dalaj say REL ‘I will tell Mikita how Dalai is talking [about him]’.

!"#$%&''()

30 Roman Shapiro

(58) Mikita osobo delo; za tibi osobo delo. Mikita special a&air TOPIC you special a&air ‘Mikita is one thing; you are another thing [so please let me buy some silk

for you]’.

All agents/topics of the dialogue are people. !e description of 19th century European clothes contains both examples of animate and inanimate topics. Za is used only with animate topics, with one exception, the already quoted sentence:

(59) Tolika za direcheki odina beleshiki kusoka perevezhi esa. only TOPIC holes one white piece tie up REL ‘Only in the hole, they tie up a white piece [of cloth]’.

!is exception could be explained by the suggestion that za is a preposition mean-ing ‘behind’ here (which is indeed the basic meaning of the Russian preposition za): ‘behind the hole they tie up a white piece’. However, this hypothesis does not work with the isolated sentence about Russian soldiers quoted by Cherepanov, as it would be really far-fetched to say that the soldiers were wearing something (themselves?) behind the cloth:

(60) Za vasha soledaza za nasha voseqanaka ponosi bylo, TOPIC your soldier TOPIC our cloth wear PAST

kako mozheno bylo za evo pokoloti. how possible PAST TOPIC he beat ‘Your soldiers were wearing our cloth, how could it be that they were beaten/

defeated’.

We would argue that in Cherepanov’s version of CPR za is always an agent/top-ic marker; it is never used as a preposition (the latter being its sole function in Russian). In the majority of cases it marks animate topics (the material is insuf-"cient to judge whether a distinction is made between people and animals, or between nouns and pronouns — as most animate topics are denoted by pronouns in the data). It sometimes marks inanimate topics, probably when they are empha-sised. On the other hand, it is omitted with secondary animate topics.

Another interesting question is the etymology of the za marker. It is di#cult to "nd any semantic or functional connection with the Russian preposition za ‘behind’, ‘instead of ’. A possible source is the Mongolian ‘approval particle’ zaa (Todaeva 1951: 166), which could be translated as ‘OK’, ‘alright’, ‘"ne’ etc.:

(61) Zaa düü min’ medev üü? OK brother my understood QUEST ‘OK my brother, did you get it?’


CPR may have used this phrase-initial particle as a politeness marker (polite-ness is a prominent feature of CPR, especially Cherepanov’s version, see under Vocabulary), which was naturally associated with animate topics: topics also tend to be phrase-initial, and politeness is usually expressed towards people. Mongolian also possesses a topic marker bol (Hammar 1983: 45–48), which obviously may have exerted only structural in$uence.

5.2.2 Location/direction markersDeictic adverbials can be used to mark location/direction in some other varieties of CPR:

(62) Xodim s’uda ixin’ dom. walk here their house ‘He came to their house’. (Bikin)

!e following example demonstrates a special case, when a deictic tam (‘there’) adverbial not only marks direction, but also acts as a link between two fragments by showing that all the roles mentioned in them pertain to the same action:

(63) Ja sor!f!taroj goda, ja pashola Perpidz’an byla, ja nachanika tam pashola. I forty-second year I went Birobidzhan PAST I boss there went ‘In 1942 I went to Birobidzhan to the boss’. (Kukan)

5.2.3 Subject/object markersIn the Kukan CPR subject and object can be marked with ivo < R ego ‘him’:

(64) !ta takoj esi, takoj ivo kitajca this such was such he Chinese ‘!ere was such a Chinese person’. (Kukan)

(65) Viti mama delaj ivo um’ai pampushka, poudza ivo Vitya mama do he can bun baozi he. ‘Vitya’s mother knew how to make buns, baozi’. (Kukan)

A similar example can be found in Shprintsyn’s records (archive):

(66) Pxusun ivo taza mynoga. Pxusun he Taz many ‘!ere are many Taz people in Pxusun’ (here ivo probably marks the topic

Pxusun).

!"#$%&''()

32 Roman Shapiro

5.3 NP word order

!e attribute precedes its headword in CPR, following Chinese, Mongolian, and Manchu word order, as in the following examples:

(67) Mongolian (Svantesson 2003: 164–165): bidnij uulzsan ter sajhan zaluugaas č we.GEN meet.PFV that beautiful young.man.ABL FOC ‘even from that beautiful young man that we have met’

(68) Manchu (Möllendorf 1892: 13): a. Nikan mudan Chinese (Kukan) pronunciation ‘Chinese pronunciation’ b. mutere baita can.be.done thing ‘a thing that can be done’

Russian word order rules for NPs are more complicated: agreed adjectival attri-butes (i.e. attributes that follow their headword in gender, case and number) are usually placed before a noun, whereas non-agreed attributes (typically nouns in the genitive case) are placed a%er it. Any word can function as a modi"er if placed before a noun in CPR,32 as the following examples illustrate:

t’oply shetany ‘warm trousers’, kitajsa pam!pushika ‘Chinese bun’, Viti mama ‘Vitya’s mother’, brata syna reb’atiseka ‘children of my brother’s son [i.e. of my nephew]’, ni sadila zhen’shen’ ‘non-planted [wild] ginseng’ (Kukan), malen’ki kurema ‘small jacket’, sukona xalaza ‘woollen robe’, vorotnika rubaxa ‘shirt collar’, poigraj l’udi ‘playing person’, sipixu l’udi ‘sleeping person’ (Cherepanov 1853), balishoy v!jdalo ‘big bucket’, dylugoj laoboda ‘another job’, niloj badzin’ke ‘rotten shoes’, kurica jajcy ‘hen’s eggs’, zhyledza daroga ‘railroad’ [a calque of both R zheleznaja doroga and C 铁路tie3lu4, literally ‘iron road’], dzilibichi dzavodza ‘brick factory’, pabulosa gumak! ‘cigarette paper’, lisyda zhyledza ‘sheet iron’ (Shprintsyn 1968).

!is structure can also be used to form idiomatic compounds:

a. calqued from Chinese: sérca-kriučóka [heart-hook], ‘evil’ (cf. C 钩心 gou1xin1 ‘hook heart = evil, treacherous’33 (Jabłońska 1957), seredece-shilo [heart-awl],

32. CPR and English word order is identical in the following set of examples, so English transla-tions may serve as glosses.

33. !e CPR word order is also possible in Chinese: 如果没有内奸, 心钩怎么可能顺利藏进密室? ru2guo3 mei2you3 nei4jian1, xin1gou1 zen3me ke3neng2 shun4li4 cang2jin4 mi4shi4? ‘If there was no mole, how could treachery [heart hook] sneak into the secret room?’ (retrieved


‘cruelty’ (same as previous, cf. C 钩 gou1 ‘sew with large stitches, weave with a hooked needle’, 钩针 gou1zhen1 ‘hooked needle’), jazyka-m’oda [tongue-honey], ‘eloquence’ (cf. C 甜言蜜语 tian2 yan2 mi4 yu3 ‘sweet words and hon-eyed phrases’,34 ruka-sapogi [hand-boots], ‘gloves’ (cf. C 手袜 shou3 wa4 ‘hand socks = gloves’) (Cherepanov 1853);

b. newly-coined: uma-konechailo [mind-ended], ‘insanity’, poceluj-parezedenika [kiss-holiday], ‘Easter’ (cf. the Russian custom of kissing each other three times on both cheeks during Easter) (Cherepanov 1853).

5.4 Word order in questions

Word order in questions is the same as in statements, just like in Chinese. As for Russian, word order is identical in statements and ‘yes-no’ questions, but the inter-rogative word (or the whole interrogative NP) usually takes the initial position in wh-questions.

Yes-no questions (69) a. CPR: Za vashe zh’onusheki mes’aca posidi esa? TOPIC your wife together sit REL ‘Do you sit together with your wives?’ (Cherepanov 1853) b. Chinese: 你们跟你们的老婆坐在一起吗? Ni3men gen1 ni3mende lao3po zuo4zai4 yi1qi3 ma you with you.POSS wife sit.COMPL together QUEST? ‘Do you sit together with your wives?’

我们跟我们的老婆坐在一起。 Wo3men gen1 wo3mende lao3po zuo4zai4 yi1qi3. we with we.POSS wife sit.COMPL together. ‘We sit together with our wives’. c. Russian: Vy sidite vmeste s zh’onami? you sit together with wives ‘Do you sit together with your wives?’

from www.ad2k.com/documents/view_694_7.html). !e data in this and next footnotes further con"rm that CPR calqued Chinese set phrases here.

34. Cf. also with the CPR order and the tongue image: 舌密笔香 she2 mi4 bi3 xiang1 tongue honey brush fragrant ‘both speak and write well’, 舌密嘴甜 she2 mi4 zui3 tian2 tongue honey mouth sweet ‘sweet words’, 舌密腹剑 tongue honey belly sword ‘a honeyed tongue with a heart of gall / an iron "st in a velvet glove’.

!"#$%&''()

34 Roman Shapiro

My sidim vmeste s zh’onami. we sit together with wives ‘We sit together with our wives’.

Wh-questions (70) a. CPR: Za tibi kakoj Pechineseki manera sholeka zakazyvaj budu? TOPIC you which Beijing manner silk order FUT ‘What kind of Beijing-style silk will you order?’ (Cherepanov 1853) b. Chinese: 你想订哪种北京绸子? Ni3 xiang3 ding4 na3zhong3 Bei3jing1 chou2zi? you want order which.CL Beijing silk? ‘What kind of Beijing-style silk will you order?’

我想订这种北京绸子。 Wo3 xiang3 ding4 zhe4zhong3 Bei3jing1 chou2zi I want order this.CL Beijing silk. ‘I will order this kind of Beijing-style silk’. c. Russian: Kakoy ty zakazhesh’ pekinskij sh’olk? which you order Beijing silk?

Or alternately: Kakoj pekinskij sh’olk ty zakazhesh’? which Beijing silk you order? ‘What kind of Beijing-style silk will you order?’

Ya zakazhu takoj pekinskij sh’olk. I order such Beijing silk. ‘I will order this kind of Beijing-style silk’.

Unlike Chinese, there are no interrogative particles in CPR. However the Chinese alternative strategy can be used to form both ‘yes-no’ and ‘wh-’ questions:

(71) Tebe dumai mnogo-malo? you think much-little? ‘How long are you going to think?’ (or ‘How much do you think [it costs]?’)

(Vrubel 1931)

(72) Pravda, ne pravda? true NEG true? ‘Is it true?’ (Kukan)


5.5 Postpositions and prepositions

Some CPR varieties use postpositions, which is common in Chinese, Mongolian, and Tungusic languages, but alien to Russian:

(73) Sobuka nizu zhivi. hill below live ‘I live at the foot of the hill’. (Shprintsyn 1968)

On the other hand, prepositions are possible in CPR:

(74) Na ruka beleneki kusoka perev’azyvaj esa. on hand white piece tie.up REL ‘!ey tie up pieces of white cloth on their wrists’. (Cherepanov 1853)

5.6 Comparatives

Comparative contsructions can be similar to constructions with postpositions: the comparative adinaka/odinakovyj can be placed a%er an NP. Cf.:

(75) a. Mongolian (Todaeva 1951: 175): Minij nöhöd nadtaj adil martašgüj ügee bajartajhelne. my friends me like unforgettable words happily.pronounce ‘My friends, like me, happily pronounce unforgettable words’. b. Chinese (Yingyong 2000: 1377): 像今天这样的事，已经 xiang4 jin1tian1 zhe4yang4de shi4 yi3jing1 like today such.POSS situation already

发生不止一次了。 fa1sheng1 bu4 zhi3 yi1 ci4 le happen not only one time INCP ‘Situations like this one today already happened more than once’.

(76) Saxala adinaka (Shprintsyn archive) sugar like ‘[sweet] as sugar’

(77) Ran’she kak zdes’a odinakovyj (Kukan) before like here same ‘!ere it used to be just like here’.

In the second example the word odinakovyj looks more like its Russian proto-type and the Russian comparative word kak is used before the NP (as in Russian).

!"#$%&''()

36 Roman Shapiro

However the whole construction still follows the Chinese model (NP1跟NP2一样; NP1 gen1 NP2 yi1yang4) rather than the Russian one (NP1 kak NP2).

CPR possesses a more typologically Russian construction35 as well, (v)s’oravno NP:

(78) Ego s’o ravno l’udi he same people ‘!ey [boars] are like people’. (Arsenyev 1983)

In Cherepanov’s CPR the comparative adali is used instead of v’o ravno in the same structure:

(79) Za nashi zhenusheki adali boga podumaj esa TOPIC our wife as.if god think REL ‘We think our wives are like goddesses’.

6. Vocabulary

6.1 Direct borrowings from languages other than Russian

!e dictionary in Perekhvalskaya (2008), combined from various sources, contains 680 lexemes. CPR vocabulary is predominantly of Russian origin.36 Borrowings from Chinese usually have synonyms of Russian origin; Jabłońska (1957) claims that the Russians tried to use more ‘Chinese’ words, and the Chinese the other way round. Although Shprintsyn (1967) disagrees with her, his data support her state-ment, as in (78) below (Shprincyn archive):

(80) Russian man: Lapyr ju,37 madamu (<R) imeesh’(<R)? wife have madam have ‘Are you married?’

Chinese man: Babushka esi. old.woman/granny have. ‘I am married’.

35. However the Russian construction with this comparative is NP1 vs’o ravno chto NP2.

36. Shprincyn (1967) thought that a pidgin based on Chinese vocabulary also existed in Russia. However his own and other authors’ (Vrubel 1931, Jabłońska 1957) data show that most pidgin sentences either use individual Chinese words on the Russian lexical ‘background’ or, even more o%en, do not use them at all.

37. Lapyr < C 老婆儿 lao3por ‘wife, old woman’; ju<C 有you3 ‘have’.


It is worth noting that babushka never means ‘wife’ in Russian; moja staruxa (‘my old woman’) can be used instead.

Table 3 contains some other examples of Russian/Chinese synonyms in CPR.

Table 3. Russian/Chinese synonymsFrom Chinese From Russian Meaning1. Bue<不要 bu4 yao4 Ne nada< ne nado ‘must not, need not’2. Chego< 这个zhe4 ge Eda<eto ‘this’3. Igejan<一个样yi1 ge yang4 Odinaka <odinakovyj like, identical4. Kanxodi<干活的gan4huo2de, sup-ported by CPR xod’a ‘(Chinese) hired hand, pedlar’ <R hodit’ ‘walk around’ and C 伙计huo3ji ‘hired hand’

Laobodaj < rabotat’, supported by Chinese 劳 lao2 ‘work’

‘work’

Most CPR words are easily etymologized. Comments on more complicated cases are suggested below.

Adali ‘like, similar’ (Cherepanov 1853) < Mongolian adali ‘equal, similar’ (Mudrak).Bichi, bichixu ‘write’ (Cherepanov 1853) < Mongolian biči(h) ‘write’ (Linguamongolia).Fanguli ‘topple over, broken, spoilt’ (Shprintsyn archive) < C 翻过来 fan1guo4lai2 ‘topple over’.

An example of the latter appears in (81):

(81) Derzhi krepko, a_to ego fanguli hold fast otherwise he topple.over ‘Hold it fast, or it is going to topple over’. (Shprincyn archive)

Kantrami ‘behead, kill’ (Shprincyn)< C 砍头儿嘛 kan3 tour2 ma/ cut head QUEST ‘behead’ (etymology by V.M. Alekseev, from Shprincyn 1967). !is etymology is supported by examples: ‘Zhang Zuolin digs out the grain and then does kantrami (cuts o& the head) to its owner’ (Voronovich 1952).

(82) Moja kantrami teb’a seki-seki my kantrami you cut-cut ‘I will slash you to death’. (Sprincyn archive)

(83) Kitajska dum lusyki gavali kantrami ivo ubila Chinese think Russian say kantrami he cut ‘!e Chinese think that ‘kantrami’ is the Russian for ‘cut’ ’ (Shprintsyn

archive).

!"#$%&''()

38 Roman Shapiro

Cf. a more general meaning:

(84) Kantrami tebe sdelajut kantrami you will.do ‘!ey will kill you’. (Veresaev 1986)

(85) komunistaj policij kantrami tebe mal-mala budet delat’ communist police kantrami you little-little will do ‘Communist police will kill you’. (Bulgakov 1990)

Karabchi ‘steal’ (Shprincyn) <R karabchit’ ‘steal, take bribes’ (Dahl 1863–1866). !is word is either originally Russian (Dahl 1863–1866) or a borrowing from Tajik (< karabčuk ‘thief ’, Trubachev in Vasmer 1964–1973) or a Turkic language (<qaraq-čy ‘robber’, Anikin 2000). It was widespread in the Caucasus pidgin, as (86) shows:

(86) tvoja b’jot, nasha vasha korobchit your beat our your korobchit ‘you beat us, and we steal from you’. (Tolstoy 1979)

A similar usage is found in Eliseev. !is word is found in modern Russian, includ-ing Siberian dialects (Siberian forum) and football slang (Football forum).

Kurema < R kurma — a kind of short overcoat, worn by Russians in the Russian Far-East (primarily by a Christian group of Old Believers, see Selischev 1920). !is piece of clothing (and the word) was probably borrowed from the Mongolians, who borrowed it from the Manchurians (see Berelekhis, Frid). Dahl explains kurma as ‘a kind of fur-coat on the Chinese border’.

Pil’uli ‘beating’ (Shprintsyn archive) < R pil’uli ‘pills’ probably confused with R bili ‘beat’ past plural + perfective su#x le. !e same meaning is used in Shikuts (1909) in a CPR context and in Kondratyev (1995) in a Russian context (probably a borrowing from CPR). A CPR derivation of the Russian pil’uli in its original meaning ‘pills’ was present in one of the ‘Russian’ textbooks (Shprincyn archive) as 必留克 bi4liu2ke4.

Pomaniza < R manishka ‘dicky, false shirt-front’; the verbal pre"x po- shows that it is associated in CPR with a verb < R manit’ ‘lure, decoy’. !is etymology is probably valid for Russian (see Vasmer 1964–1973).

Shaletaj-boletaj ‘higgledy-piggledy’ (Cherepanov 1853) cf. R shaltaj-boltaj. !e Russian word shaltaj-boltaj means ‘idle, aimless’ (Ozhegov 1990), ‘good-for-noth-ing’ (Efremova 2000), or ‘rubbish’ (Ushakov 1935–1940). Dahl attributes shaltaj-baltaj to Orenburg and Siberian Russian dialects (which interacted with CPR) and explains it as ‘idle talk, rubbish’. Dahl also records shaltat’, shaltyxat’ ‘babble, prat-tle (of children)’ (Pskov and Tver dialects). !e second part — boltaj (imperative;


in"nitive boltat’) — means ‘chatter’ in standard Russian. If Dahl’s explanation is correct then shaltaj-boltaj must have been borrowed into CPR from Russian, and not vice versa, as Pskov and Tver regions and quite far from the pidgin zones. On the other hand, it could be explained as ‘an argotic formation from the truncated boltat’ with the sha- pre"x added’ (Vasmer 1964–1973). Reduplication of this kind is present in many languages, including Russian and Chinese (cf. R tary-bary ‘chat-ter’, C 唠叨 lao2dao ‘be garrulous, chatter’, 混沌 hun4dun4 ‘chaos; muddle-headed and ignorant’). !erefore shaltaj-boltaj could have been formed both in CPR and in Russian. However, the use of imperatives not in their regular meaning is typical of CPR and rare in Russian. In modern Russian Shaltaj-Boltaj is mainly used as the translation (originally by Samuil Marshak) of Humpty Dumpty, which is quite popular in Russia because of Lewis Carroll’s "rough the Looking-Glass.

Shango ‘good’ (Schuchardt 1884) < C 上好 shang4hao3 or 上高 shang4gao1 ‘very good’. However this word, as well as kantrami, was perceived by the Chinese as a Russian one (Shprincyn archive).

Soli-povesi ‘die’ (Cherepanov 1853)< C 死了si3le or si3liao3 ‘dead’ (?)+ R povesit’ ‘hang’.

6.2 Calques

Chinese also in$uenced CPR vocabulary by way of calques. Structural calques (seredece-kr’uchok etc.) have been discussed in Section 5.3. Consider several ad-ditional examples:

(87) Ni nada setisn’aja! Ni nada setisn’aja! NEG need be_shy NEG need be_shy ‘Welcome!’ (Kukan)

Compare R ne nado stesn’at’s’a ‘don’t be self-conscious’, in$uenced by the C phrase in (88):

(88) 别客气 bie2 ke4qi4 NEG polite/self-conscious ‘Welcome’ (lit. ‘Don’t be [too] polite/self-conscious’)

!e Chinese phrase (as well as the CPR example) is typically used to o&er something to a guest. However, the Russian sentence may be used only in its literal meaning.

(89) Viti mama Vitya (speaker’s son) mummy ‘Vitya’s mother’ (Kukan)

!"#$%&''()

40 Roman Shapiro

!e formula ‘my child’ + ‘mother’ is used to mean ‘my wife’, as is typical of Chinese, for example in (90):

(90) 孩子他娘 hai2zi ta1 niang2; child his mother ‘my child’s mother = my wife’

Chivo-chivo ‘whatever’(Shprincyn) < R chego ‘what’ in$uenced by C 什么什么shen2me-shen2me ‘what-what/‘whatever’; mnogo-malo ‘how many’ (Vrubel 1931) < R mnogo ‘much’ + malo ‘little’ in$uenced by C 多少 duo1shao / much little / how many.

Semantic calques include: sonce ‘day’ (Schuchardt 1884)< R solnce ‘sun’ cf. C 日ri4 ‘sun, day’; babushka ‘wife’ (Shprincyn) < R babushka ‘grannie/old woman’, cf. C 老婆 lao3po ‘old woman/wife’; ivo ‘I’ < R ego ‘he’. An example of the latter appears in (91):

(91) Ivo paruski ni gavalila he in.Russian NEG speak.PAST ‘I didn’t speak Russian’. (Kukan)

Compare C 人家 ren2jia ‘they, he, I’, probably also under phonetic in$uence of C 我 wo3 ‘I’. Another example of a calque is the use of ise38 as an optional plural marker (< R vse ‘all’):

(92) Is’e ukalainza s’uda prish’ola all Ukranian here came ‘Ukranians came [to live] here’. (Kukan)

Cf. the use of C 都 dou1 ‘all’:

(93) 老师来了 lao3shi1 lai2 le teacher come INCP

(94) 老师都来了 lao3shi1 dou1 lai2 le teacher all come INCP.

38. It is not always easy to di&erentiate is’e from is’o (<R eq’o ‘more/still’ or R vs’o ‘all [neuter]’), as Cyrillic CPR sources, according to the Russian spelling convention, tend to use e for both e and ё.


Sentence (94) is more likely to be understood in the plural (‘Teachers came’), whereas (93) in the singular (‘!e teacher came’). A similar phenomenon is found in Taimyr pidgin (Xelimsky 2000: 385, questioned in Stern 2005a).

A very interesting kind of Chinese in$uence on CPR is ‘diminutive polite-ness’ (see Jurafsky 1996 for cross-linguistic typology), which is usually expressed in Chinese by 一点 yi1 dian3, 一下 yi1 xia4, 一会 yi1 huir4, and 稍微 shao1 wei1 ‘a little, a while’. In CPR ‘diminutive politeness’ is expressed:

a. lexically, as in the following:

(95) Sichasa malo-malo laobotaj now little-little work ‘I work just a little now’. (Vrubel 1931)

Or even as in (96):

(96) Kantrami tebe mal-mala budet delat’ kill you little-little will do ‘[!ey] will kill you a little’. (Bulgakov 1990)

b. with a#xes. !e po- verbal pre"x, present in practically all verbs in Cherepanov (1853), this pre"x typically means ‘for a while’ in Russian: CPR poguli ‘walk’ < R pogul’at’ ‘walk for a while’ cf. gul’at’ ‘walk’ etc. Cherepanov’s pidgin also widely uses diminutive su#xes (in bold): beleneki ‘white’, cf. R belyj; zh’onusheki ‘wife’ cf. R. zhena; r’umasheka ‘wineglass’ cf. R r’umka. !ese ways of expressing ‘diminutive politeness’ exist in Russian as well, however in Cherepanov’s CPR they are lexicalized and used far more o%en.

7. Conclusions

!e following table summarizes the main features of CPR in comparison with the languages involved in its formation. Shading shows the degree of a#nity.

It can be seen that Russian most in$uenced CPR in terms of lexicon and, to a lesser extent, morphology; Chinese in$uence in CPR can be seen in terms of phonotactics and, to a lesser extent, syntax; interestingly, Mongolian and Tungusic in$uence is also found in CPR in certain areas of syntax.

!is paper is much indebted to the previous scholars in the "eld, especially for their data. It however contributes its mite toward the study of Chinese Pidgin Russian. To the best of our knowledge, this is the "rst complete account of CPR that describes all its key aspects (history, phonology, morphology, syntax, and vo-cabulary) and takes into consideration the in$uence of both its ‘parent’ languages,

!"#$%&''()

42 Roman Shapiro

i.e. Russian and Chinese, as well as other less obvious in$uences from Mongol and Tungusic languages. It uses data from a wide range of sources, some of which are rarely accessible to Western linguists (especially Shprintsyn’s archive). On the analytical side, this study proposes a hypothesis on tone inheritance in contact languages and provides some arguments in its support. It suggests an opposition between verbs and non-verbs as the two major morphological classes of CPR. It o&ers new interpretations of "nal particles (taking into account their Chinese

Table 4. CPR features and originsFeature CPR Russian Chinese Mongolian and

TungusicConsonantclusters

avoided frequent Impossible within syllables;restricted atsyllable boundaries

restricted within syllables;frequent at syllableboundaries

Tone absent absent present absentParts of speech Verbs vs. non-

verbs (etym. adjectives and nouns merged into non-verbs); Russian markers used

A well-devel-oped system; ad-jectives grouped with nouns

A well developed system; adjec-tives grouped with verbs

A well-devel-oped system; ad-jectives grouped with nouns

Role coding Reduced, word order and syntac-tic words

cases Word order and syntactic words

cases

personal / pos-sessive pronouns

merged di&erent similar similar

Basic word order Reduced SOV SVO SVO SOVNP word order Head-last Head-last in

agreed NPs Head-"rst in non-agreed NPs

Head-last Head-last

Locatives Prepositions and postpositions

prepositions postpositions postpositions

Sentence-"nal TMA markers

Present, mostly similar to auxil-iary verbs

absent Present, di&erent from auxiliary verbs

Auxiliary verbs

In$uence on CPR vocabulary

– Source of most CPR vocabulary

Some words and lexical structures

sporadic


counterparts) and a survey of question structure in CPR. Finally, it gives new trans-lations and etymologies for ‘di#cult’ CPR words and phrases. !e future research plan includes an in-depth typological comparison with other contact languages.

References

Aleksandrov, A. 1884. (Александров А.) [Maimachin dialect] Маймачинское наречие // Русский филологический вестник. Т. XII, №3, Варшава, 160–163.

Alekseev, M.P. 1968. (Алексеев М.П.) [Foreign word vocabularies in a 17th century Russian ABC] Словари иностранных слов в русском азбуковнике XVII в. Ленинград.

Anikin, A.E. 2000. (Аникин А.Е). [Etymological dictionary of Siberian Russian dialects: Uralic, Altaic and Paleo-Asiatic loanwords] Этимологический словарь русских диалектов Сибири: Заимствования из уральских, алтайских и палеоазиатских языков. Новосибирск, Наука.

Ansaldo, Umberto. 2009. Contact languages: Ecology and evolution in Asia. New York: Cambridge University Press.

Arsenyev, V.K. 1983. (Арсеньев В.К.) [In the Ussuri region]. По Уссурийскому краю. Дерсу Узала. Москва, Правда.

Avery, Martha. 2003. !e Tea Road: China and Russia meet across the Steppe. 五洲传播出版社。Beijing.

Belikov, V.I. 1997. (Беликов В.И.) [Russian Pidgins] Русские пиджины // Малые языки Евразии: социолингвистический аспект. Москва, Издательство МГУ, 90–108.

Berelekhis, O., Frid, M. (Берелехис O., Фрид M.) [Mongolian clothes] Монгольский костюм. http://orkclub.ru/biblioteka/bibri/bibrirekvizit/304-masterskaja-igrovojj-i-teatralnojj-ode-zhdy.html)

Bikin: Collected in 1985 and 2004 from the Udihes in Krasny Yar village, Pozharskiy district, Primorye, Russia. Perekhvalskaya 2008.

Broch, Olaf. 1927. Russenorsk. Archiv für Philologie, 41, 209–262.Broch, Olaf. 1930. Russenorsk Textmateriale. Maal og Minne, H. 4, Oslo, 113–140.Broch, Ingvild & Jahr, Ernst. H. 1981. Russenorsk — et pidginspråk i Norge/Tromsø-studier i

språkvitenskap 3. Oslo.Broch, Ingvild & Jahr, Ernst H. 1984. Russenorsk: A new look at the Russo-Norwegian pidgin

in northern Norway. In P.S. Ureland & I. Clarkson, Scandinavian language contacts, 21–65. Cambridge: Cambridge University Press.

Broch, Ingvild & Jahr, Ernst H. 1990. Russenorsk: !e Russo-Norwegian pidgin. New "ndings. Tromsø linguistics in the eighties. Tromsø-studier i språkvitenskap, 11. Oslo: Novus.

Bulatova, N. Ya. 1999. (Булатова Н.Я.) [Sakhalin Evenki language] Язык сахалинских эвенков. Б&К, Санкт-Петербург.

Bulgakov, M.A. 1990. (Булгаков М. А.) [Zoyka’s $at] Зойкина квартира // Cобрание сочинений в пяти томах, том третий. Москва, Художественная литература.

Chao, Yuen Ren. 1968. A Grammar of Spoken Chinese. Berkeley: University of California Press.Cherepanov, S.I. 1853. (Черепанов С.И.) [Kyakhta Chinese dialect of Russian] Кяхтинское

китайское наречие русского языка // Известия Академии наук по отделению русского языка и словесности. Т.2.

!"#$%&''()

44 Roman Shapiro

Dahl, V.I. 1863–1866. (Даль В.И.) [Explanatory dictionary of the living Great-Russian lan-guage] Толковый словарь живого великорусского языка. http://dic.academic.ru/con-tents.nsf/enc2p/

Demidova, N.F. & Miasnikov, V.S. 1966 (Демидова Н.Ф., Мясников В.С.) [First Russian diplo-mats in China] Первые русские дипломаты в Китае. Москва.

Demidova, N.F. & Miasnikov, V.S. 1969. (Демидова Н.Ф., Мясников В.С.) [Russian-Chinese relations in the 18th century] Русско-китайские отношения в XVII в. Материалы и документы. Т. 1. 1608–1683. М. Наука. 1969. http://www.vostlit.info/Texts/Dokumenty/China/XVII/1600-1620/Rus_kit1/

Efremova, T.F. 2000. (Т. Ф. Ефремовa.) [New expanatory derivation Russian dictionary] Новый толково-словообразовательный словарь русского языка под ред. Т. Ф. Ефремовой. http://dic.academic.ru/contents.nsf/efremova/

Eliseev, F.I. (Елисеев Ф.И.) [With Khoper soldiers] С хоперцами. http://www.dk1868.ru/history/s_xoper2.htm

Football forum: forum.soccergames.ru/archive/index.php/t-914.htmlFlug, K.K. 1935 (Флуг К.К.) [A short survey of the non-Buddhist Chinese manuscript col-

lection of the Oriental Studies Institute, USSR Academy of Sciences] Краткий обзор небуддийской части китайского рукописного фонда Института Востоковедения Академии наук СССР // Библиография Востока, вып. 7. 1934. Москва-Ленинград, Издательство Академии наук, 87–92.

Hammar, Lucia B. 1983. Syntactic and pragmatic options in Mongolian — a study of bol and n'. Ann Arbor: Indiana University.

Hua E chuyu 1927: Chinese and Russian conversation for beginners. 华俄初语. Первоначальные китайско-русские разговоры. 6 издание. Харбин.

Jabłońska Alina. 1957. Język mieszany chińsko-rosyjski w Mandżurii. Przegląd Orientalistyczny 21, 157–168.

Jurafsky Daniel. 1996. Universal tendencies in the semantics of the diminutive. Language 72:3, 533–578.

Kondratyev, V.L. 1995 (Кондратьев В.Л.) [Regards from the battle-front] Привет с фронта. Повести и рассказы. Москва, Художественная литература. http://militera.lib.ru/prose/russian/kondratyev3/index.html

Korsak, A.K. 1857. (Корсак А.К.) [Historical and statistical survey of Russian-Chinese trade relations] Историко-статистическое обозрение торговых сношений России с Китаем. Казань.

Kukan: Collected in Kukan village (Khabarovsk district, Russia) in 1990 from a Chinese, who had lived in Russia since 1942, and his Russian-speaking son. Perekhvalskaya 2008.

Kyakhta. 1990: [Kyakhta. Historical and cultural monuments] Кяхта. Памятники истории и культуры: путеводитель. Москва.

Lewicki, Marian. 1949. La langue mongole des transcriptions chinoises du XIV-e siècle. Le Houya-yi yi-yu de 1389. Wroclaw.

Li, Boya 2006. Chinese "nal particles and the syntax of the periphery. !e Leiden University Centre for Linguistics. https://openaccess.leidenuniv.nl/bitstream/1887/4393/1/!esis.pdf

Lim, Lisa. 2009. Revisiting English prosody: (Some) New Englishes as tone languages? English World-Wide 30:2, 218–239(22). Amsterdam: John Benjamins.

Linguamongolia: http://www.linguamongolia.co.uk/dict1.htmlLufrano, Richard John. 1997. Honorable merchants: commerce and self-cultivation in late-im-

perial China. University of Hawaii Press.


Lü, Shuxiang 2007: 吕叔湘. [800 words of Modern Chinese] 现代汉语八百词. 北京, 商务印书馆.Maksimov, S.V. 1871 (Максимов С.В.) [In the Orient. A Journey to the Amur river. Travels

and memoirs] На Востоке. Поездка на Амур: Дорожные заметки и воспоминания. Санкт-Петербург.

Masini, Federico. 1993. !e formation of Modern Chinese lexicon and its evolution toward a national language. California University Press.

Möllendor&, Paul Georg von. 1892. A Manchu grammar: With analysed texts. Shanghai.Moskovskiy Telegraf. 1831: О. Иоакинф (Бичурин) (?). [A Letter from Kyakhta] Письмо из

Кяхты // Московский телеграф №21 (141). С.5.Mudrak, O.A. Mongolian Etymological Database. Available online: http://starling.rinet.ru/cgi-

bin/query.cgi?root=con"g&morpho=0&basename=\data\alt\mongetNeumann, Günter. 1966. Zur chinesisch-russischen Behelfssprache von Kjachta. Die Sprache

12, 237–251.Nichols, Johanna. 1980. Pidginization and foreigner talk: Chinese Pidgin Russian. In E.C.

Traugott, R. Labrum, & S. Shepherd (eds.), 4th International Conference on Historical Linguistics, 397–407. Amsterdam, John Benjamins.

Nichols, Johanna. 1986. !e bottom line: Chinese pidgin Russian. In W. Chafe, J. Nichols (eds.), Evidentiality: "e linguistic coding of epistemology, 239–257.

Oglezneva, E.A. 2007a. (Оглезнева Е.А.) [Russian-Chinese Pidgin: An attempt at a socio-linguistic description] Русско-китайский пиджин: опыт социолингвистического описания. Благовещенск, 2007.

Oglezneva, E.A. 2007b. (Оглезнева Е.А.) [Russian-Chinese Pidgin in the Far East: features of functioning and linguistic speci"city] Русско-китайский пиджин на Дальнем Востоке: особенности функционирования и языковая специфика // Известия РАН. Серия литературы и языка. 2007. Том 66, № 4, 35–52.

Ozhegov, S.I. 1990 (Ожегов С.И.) [Russian Dictionary] Словарь русского языка. http://dic.academic.ru/contents.nsf/ogegova/

Pallas, Peter. S. 1776. Reise durch verschiedene Provinzen des russischen Reichs. III: Vom Jahr 1772 und 1773. St. Petersburg.

Perekhvalskaya, E.V. 2003. Quanti"cation in the Russian-Chinese Pidgin. In Pirkko Suihkonen and Bernard Comrie (eds.), Collection of papers of the Int. Symp. On deictic systems and quanti$cation in languages spoken in Europe and north and central Asia, 153–163. Udmurt State Univ., Max Plank Institute of Evolutionary Anthropology.

Perekhvalskaya, E.V. 2004. Clause structure in the Russian-Chinese Pidgin. In Bernard Comrie, Pirkko Suihkonen, and V. Solovyev (eds.), Proceedings of the Intl. Symp. “Typology of argu-ment structure and grammatical relations”, 94–97. Kazan State Univ., Max Plank Institute for Evolutionary Anthropology.

Perekhvalskaya, E.V. 2008. (Перехвальская Е.В.) [Russian Pidgins] Русские пиджины. Москва, Алетейя.

Qing E: !e complete Chinese and Russian conversations. 清俄会语俱全校正无讹. Полное собрание китайско-русских разговоров. Проверено и исправлено, без ошибок. Block-printed, no date.

Schuchardt, Hugo. 1884. (Шухардт, Гуго.) [Maimaichin dialect] Маймачинское наречие // Русский филологический сборник. Варшава, T. XII, №4. Pp. 318–320.

Selischev, A.M. 1920 (Селищев А.М.) [Transbaikalian Old Believers] Забайкальские старообрядцы. Семейские. Иркутск. http://www.krotov.info/history/20/1910/1920seli.html

!"#$%&''()

46 Roman Shapiro

Shikuts, F.I. 1909. (Шикуц Ф.И.) [A soldier’s diary during the Russian-Japanese war] Дневник солдата в Русско-японскую войну: В 2-х ч./ Под ред. В. И. Пржевалинского. СПб.: Сенат. типогр. http://militera.lib.ru/db/shikuts_"/index.html

Shishkov, V.Ya. 1989 (Шишков В.Я.) [!e Ugrium River / Gloomy River] Угрюм-река. Ставропольское книжное издательство.

Shprintsyn, A.G. 1932. (Шпринцын А.Г.) [!e Chinese-Russian dialect as analysed by S.A. Vrubel] Китайско-русский диалект в освещении С.А. Врубеля // Революция и письменность, №№ 1–2, Москва.

Shprintsyn, A.G. 1968. (Шпринцын А.Г.) [On the Russian-Chinese dialect in the Far East] О русско-китайском диалекте на Дальнем Востоке // Страны и народы Востока. Вып. 6, 86–100.

Shprintsyn archive: Shprintsyn, A.G. (Шпринцын А.Г.) [Archive] Архив. Государственная публичная библиотека им. М.Е. Салтыкова-Щедрина. Отдел рукописей, ф. 1200. СПб.

Siberian forum: http://www.solidwaste.ru/phorum/print.html&f=1&t=56&page=78Siegel, Je&. 2000. Processes of language contact: Studies from Australia and the South Paci"c.

Saint-Laurent, Quebec : Fides.Skachkov, P.E. 1966. (Скачков П.Е.) [History of 17th-18th cc. Chinese Studies in Russia]

История изучения Китая в России в 17–18 вв. Москва.Skachkov, P.E. 1977. (Скачков П. Е.) [Essays on the history of Russian sinology] Очерки

истории русского китаеведения. Москва, Наука.Stern, Dieter. 2002. Russische Pidgins. Die Welt der Slaven XLVII, 1–30.Stern, Dieter. 2005a. Myths and facts about the Kyakhta trade pidgin. Journal of Pidgin and

Creole languages 20: 1, 175–187.Stern, Dieter. 2005a. Taimyr Pidgin Russian (Govorka). Russian linguistics, 289–328.Svantesson, Jan-Olof. 2003: Khalkha. In: Janhunen 2003, 154–176.Taz: Collected in 1990 from the Taz in Mikhaylovka village, Olginskiy district, Primorye, Russia.

Perekhvalskaya, 2008.!omason, Sarah Grey & Kaufman, Terrence. 1988. Language contact, creolization, and genetic

linguistics. Berkeley: University of California Press.Timkovskiy, E.F. 1824 (Тимковский Е.Ф.) [Journey to China through Mongolia in 1820 and

1821] Путешествие в Китай через Монголию в 1820 и 1821 годах. Т. I. Saint-Petersburg.Todaeva, B.Kh. 1951 (Тодаева Б.Х.) [Modern Mongolian Grammar] Грамматика современного

монгольского языка. Изд-воa Н СССР. Москва.Tolstoy, Leo. 1979 (Толстой Л.Н.) [!e Cossacks] Казаки // Собрание сочинений в 22 томах.

Том 3. Москва, Художественная литература.Tsintsius, V.I. et al. 1975–1977 (Цинциус В.И. и др.) [Comparative dictionary of Tungusic-

Manchu languages] Сравнительный словарь тунгусоманьчжурских языков. Ленинград. http://www.altaica.narod.ru

Tsintsius, V.I. 1982 (Цинциус В.И.) [Negidal language] Негидальский язык. Наука, Ленинград.

Ushakov, D.N. 1935–1940. [Russian Explanatory Dictionary] Толковый словарь русского языка: В 4 т. / Под ред. Д. Н. Ушакова. http://dic.academic.ru/contents.nsf/ushakov/

Vasmer, Max. 1964–1973 (Фасмер, Макс.) [Russian Etymological Dictionary] Этимологический словарь русского языка. http://dic.academic.ru/contents.nsf/vasmer/

Veresaev, V.V. 1986. (Вересаев В. В.) [A doctor’s notes] Записки врача. На японской войне. Москва, Правда.


Voronovich, N.V. 1952. (Воронович Н.В.) [Russian-Japanese war] Русско-японская война. Воспоминания. Нью-Йорк.

Vrubel, S.A. 1931. (Врубель С.А.) [Russian-Chinese linguistic crossing] Русско-китайские языковые скрещения // Культура и письменность Востока, № 7–8. Москва.

Wee, Lian-Hee. 2007. Englishization as an aspect of building the Singapore identity. To appear in Englishization in Asia, edited by Kwok Kan Tam. Hong Kong Open University Press.

Winford, Donald. 2006. “Reduced syntax in prototypical pidgins”. In Ljuljana Progovac et al. (eds.), "e Syntax of Non-sententials, 283–307. Amsterdam: John Benjamins.

Wu, Guo. 2005. !e discourse function of the Chinese particle ne in statements. JCLTA 40:1, 47–82.

Xelimskij E.A. 1987. (Хелимский Е.А.) [We’ll speak Govorka] «Русский говорка место казать будем» (Таймырский пиджин) // Возникновение и функционирование контактных языков. Материалы рабочего совещания. Москва, 84–93.

Xelimskij, Evgenij A. 1996. ‘Govorka’ — the pidgin Russian of the Taymyr Peninsula Area. In Steven A. Wurm, Peter Mühlhaüsler, & Darrel T. Tryon (eds.), Atlas of languages of inter-cultural communication in the Paci$c, Asia, and the Americas. Mouton de Gruyter, Berlin.

Xelimskij, E.A. 2000 (Хелимский Е.А.) [Govorka — the Taymyr pidgin on the lexical base of Russian] Говорка — таймырский пиджин на русской лексической основе // Хелимский Е.А. Компаративистика, уралистика: лекции и статьи. Москва.

Yan, Guodong 2006. (阎国栋). [History of Russian Sinology] 俄国汉学史：迄于1917年. 人民出版社, 北京.

Yang, Jie 2007. (Ян Цзе) [Transbaikalian Manchu Pre-Pidgin] Забайкальско-маньчжурский препиджин. Опыт социо-лингвистического исследования // Вопросы языкознания. №2., 67–74.

Yingyong Hanyu Cidian 2000. [A Practical Chinese Dictionary] 应用汉语词典. 商务印书馆, 北京.

Zhong E huaben 1902: Russian and Chinese conversation.中俄话本. Книга китайско-русских разговоров. Сост. Кауфман /俄国高福满. Shanghai.

Zhong E tongsu 1913: Chinese and Russian vernacular conversations.中俄通俗话本. Книга китайско-русских простонародных разговоров. Сост. Майский / 麦斯克, 李明清.

Appendix

‘A Letter from Kyakhta’ (Moskovskij Telegraf 1831: 5):

(97) Eding liudi bolishele. one people came. ‘Someone has come’.

(98) Tibi, buliyatil[e], ya nie shenale. you,39 friend I not know. ‘I don’t know you, my friend’.

39. An oblique case, the transcription does not show which (grammatical Russian would use accusative here).

!"#$%&''()

48 Roman Shapiro

(99) Tibi duwal[e] gaohuda huodi? you.OBL commodity when walk? ‘When are your goods arriving?’

(100) Wuchila bolishele. yesterday came. ‘[!e goods] arrived yesterday’.

(101) Ti gekai duwal[e] yesi? you.NOM which commodity is? ‘Which goods do you have?’

(102) Ya xiake duwal[e] yesi. I.NOM any commodity is ‘I have all kinds of goods’.

(103) Ti gekai duwal[e] nadu? you.NOM which commodity need ‘What kind of goods do you need?’

(104) Edika duwal[e] ya nia[n] nadu. this.GEN.SG[?] commodity I not need ‘I do not need this commodity’.

(105) Hulaoshi duwal[e] dai. good commodity give. ‘Give/sell me good things’.

(106) Caina gaowuli. price say. ‘Tell me the price’.

(107) Lishika na buluoxi. excess NEG ask. ‘Don’t ask to much [money]’.

(108) Eding duwa silaowu. one two word ‘Say it in one or two words / don’t beat about the bush’.

(109) Daliaoka gaowuli. far say ‘You have wandered too far from the subject’.

(110) Buliaoma caina gaowuli. straight price say ‘What’s the price, give it to me straight’.

(111) Ti mala-mala bulibeiwei. you.NOM little little add ‘Raise the price a little’.


Cherepanov (1853):Features of old Russian spelling in the pidgin transcription are ignored in our romanized trans-literation, as they had no phonetic relevance for the 19th century Russian, and were most prob-ably used just to make the trancription closer to the standard Russian usage of the time.

A conversation between a Chinese merchant (C) and a Russian merchant’s wife (R):

(112) C: Ilisandera, za tibi kaka podumaj, za moja prish’ol esa? Alexandra, TOPIC you how think, TOPIC my come.INCP REL. ‘Alexandra, why do you think I have come?’

(113) R: Poguli esa, gospodine Dalaj. walk REL, Mr. Dalai. ‘You are having a walk, Mr. Dalai’.

(114) C: Poguli malo; za moja podumej, kako vasha zakona xoroshaneki: walk little; TOPIC my think how your law good

zh’onusheki mes’aca pozhivu esa. wife together live REL ‘Not only having a walk; I think your custom is so good: [husbands] live together with

their wives’.

(115) R: Pocheto za vashi zh’onusheki s’uda poxodi nitu? why TOPIC your wife here come NEG ‘Why don’t your wives come here?’

(116) C: Zakona nitu; za evo pomeshaj potorgovaj budu. law NEG TOPIC he interfere trade[verb] FUT ‘!ere is no such custom; they will interfere with the trade’.

(117) R: Kak mozheno; za nasha kakoj pomeshaj nitu? how possible TOPIC our which interfere NEG ‘Is this possible? Why is it that we do not interfere?’

(118) C: O, za vasha osoba manera; za vasha muzha zhesetoki oh TOPIC your special manner TOPIC your husband cruel

manera pozhivu esa; za zh’onusheki mes’aca posidi netu; manner live REL TOPIC wife together sit NEG

adali chuzhoj as.if stranger ‘Oh, you have a special way of life. You have husbands who live in a rude way. Wives

do not sit together [with their husbands]. Just like strangers’.

!"#$%&''()

50 Roman Shapiro

(119) R: Za vashe zh’onusheki mes’aca posidi esa? TOPIC your wife together sit REL ‘Do you sit together with your wives?’

(120) C: Neperemena, za nashi zh’onusheki adali boga podumaj esa. certainly TOPIC our wife as.if god think REL

Na tibi posemoteri, adali ledeneza kushaxu esa… on you look as.if lollipop eat REL

odina dena za tibi poxodi nitu, za moja soli-povesi budu. one day TOPIC you come NEG TOPIC my die FUT ‘Of course, we think our wives are like goddesses. Looking at you is like eating a lol-

lipop. If you don’t come [to me] one day I will die’.

(121) R: !, polnote, g. Dalaj, kako mozheno tako pogovori. eh come.on Mr. Dalai how possible so say ‘Eh, come on, Mr. Dalai, how can you say that!’

(122) C: Per’amo solovo; za moja fal’shivajla nitu; za tvoja Mikita direct word TOPIC my false NEG TOPIC your Mikita

der’ani l’udi; tol’ko r’umasheki kushaxu um’esha. rubbish person only wineglass eat can ‘Upon my word; I am not lying; your Mikita is a rotter; the only thing he’s good at is

drinking’.

(123) R: Za moja Mikita skazyvaj budu, kako Dalaj pogovori esa. TOPIC my Mikita tell FUT how Dalai say REL ‘I will tell Mikita how Dalai is talking’.

(124) C: Skazyvaj ne nado; sama podumaj mozheno; za moja tell NEG need oneself think possible TOPIC my

Pechineseki poxodi budu; za tibi kakoj Pechineseki manera Beijing go FUT TOPIC you which Beijing manner sholeka zakazyvaj budu? silk order FUT ‘No need to tell him; one could think about oneself; I will go to Beijing; what kind of

Beijing-style silk will you order?’

(125) R: Moneza nitu, g. Dalaj. money NEG Mr. Dalai ‘I don’t have money, Mr. Dalai’.


(126) C: Podozhedi mozhno. wait possible ‘I could wait’.

(127) R: Posele otdavaj nado budu. later return need FUT ‘I’ll have to pay back later’.

(128) C: Shelataj-baletaj mozheno. higgledy-piggledy possible ‘Just do it’.

(129) R: Za moja tako ni xychi. TOPIC my so NEG want ‘I don’t want to do things like that’.

(130) C: Kako xychi? how want? ‘How then?’

(131) R: Za moja xychi Mikita pogovori. TOPIC my want Mikita talk ‘I want to talk to Mikita’.

(132) C: Mikita osobo delo; za tibi osobo delo. Mikita special a&air TOPIC you special a&air ‘Mikita is one thing; you are another thing’.

(133) R: Za moja tako ne xychi. TOPIC my so NEG want ‘I don’t want to do things like that’.

(134) C: Nu, sama vol’a, posel’a zhalej budu. well oneself will later regret FUT ‘Well, it’s up to you; you are going to be sorry later’.

Two sentences about the cloth that a Russian merchant Nerpin sent from Kyakhta for the Russian militia in 1812 (to "ght Napoleon’s troops):

(135) Za nasha znasha, taka budi. Za vasha soledaza za nasha TOPIC our know so FUT TOPIC your soldier TOPIC our

!"#$%&''()

52 Roman Shapiro

voseqanaka ponosi bylo, kako mozheno bylo za cloth wear PAST how possible PAST TOPIC

evo pokoloti. he beat. ‘We knew it would be so. Your soldiers were wearing our cloth, how could it be that

they were beaten/defeated’.

An oral pidgin translation by a Chinese of his letter home about Russian clothes:

(136) Za tibi serediza budu… TOPIC you angry FUT ‘You will be angry’.

(137) Tuta vasha manera bichi esa. here your manner write REL ‘Your way of life is described here’.

(138) Odina shuba nitu, toliko odina sukona xalaza vorotenika mexa pirishivaj esa. one fur.coat NEG, only one cloth robe collar fur sew.on REL ‘Some people don’t have fur-coats, they just have fur collars sewn on cloth robes’.

(139) Za evo malen’ki kurema pode-bol’shana kurema ponosi esa; TOPIC he small jacket under-big jacket wear REL ‘they wear a small jacket [i.e. a waistcoat] under a big jacket [i.e. a tailcoat];’

(140) tuto pereza sh’oleka, zada xoleseta; here front silk back cloth ‘there front part is silk, and the back part is cloth;’

(141) sholeka pokazyvaj esa, pobol’shane kurema adali vorota ne zapiraj esa; silk show REL big jacket as.if gate NEG shut REL ‘Silk is shown o&, the big jacket is not buttoned up just like an open gate;’

(142) malen’ka kurema, sama pobol’shen sutuzha, dirocheki small jacket most big frost hole

osatavl’aj esa, za evo belin’ki rubaxa pokazyvaj nado leave REL TOPIC he white shirt show need ‘!e small jacket [waistcoat] has a low neckline, in spite of most severe frosts, as they

need to show the white shirt’;

(143) za tibi podumaj, sama.dela beleneki rubaxa ponosi esa? TOPIC you think very.a&air40 white shirt wear REL ‘Do you think they really wear white shirts?’

(144) Ojo, rubaxa ponosi toleseta, posereta, tolika za dirocheki Oh, shirt wear thick motley only TOPIC hole

40. !e Russian idiom v samom dele (lit. ‘in very a&air’) means ‘really’.


odina beleshiki kusoka perevezhi esa pomaniza nazyvaj esa. one white piece tie.up REL dicky/decoy call REL ‘Oh, they wear thick motley shirts, only in the hole they tie up a white piece [of cloth],

it is called pomaniza’.

(145) Za ego sheja palatoka ponosi esa; TOPIC he neck kerchief wear REL ‘!ey wear kerchiefs/foulards around their necks;’

(146) napereza beleneki kusoka, adali vorotinika rubaxa perev’azyvaj esa; "rst white piece as.if collar shirt tie.up REL ‘First they tie up a white piece [of cloth], like a shirt collar’;

(147) pobol’shana kuremu rukava koroteneki, na ruka beleneki big jacket sleeves short on hand white

kusoka perev’azyvaj esa: bela rubaxa obmani nado. piece tie.up REL white shirt deceive need. ‘!e sleeves of the big jacket [tailcoat] are short, white pieces [of cloth] are tied on the

wrists [as false cu&s]: one should deceive [everybody] on account of the white shirt’.

Grammar examples:Sentence-$nal particles‘Relevance’ [REL] particle esa, esi, ju (see Section 4.3.1):A. (148) Za moja Mikita skazyvaj budu, kako Dalai pogovori esa. TOPIC my Mikita tell FUT how Dalai say REL ‘I will tell Mikita how Dalai is talking’. (Cherepanov 1853)

(149) Kazeny enigi davaj esi. public money give REL ‘!ey are giving out public money’. (Shprintsyn archive)

(150) Tibe chivo gavali gavali mynoga esi? you why speak speak much REL ‘Why are you talking so much?’ (Shprintsyn archive)

B. (151) Zhenusheki mes’aca pozhivu esa. wife together live REL ‘[Husbands] live together with their wives’. (Cherepanov 1853)

(152) Za tibi podumaj sama dela beleneki rubaxa ponosi esa? TOPIC you think very a&air white shirt wear REL ‘Do you think they really wear white shirts?’ (Cherepanov 1853)

C. (153) Ilisandera, za tibi kaka podumaj, za moja prishel esa? Alexandra TOPIC you how think TOPIC my come.INCP REL ‘Alexandra, why do you think I have come here?’ (Cherepanov 1853)

(154) Tvajá tri sónca kupí jest’. your three sun buy REL ‘You bought it three days ago’. (Schuchardt 1884)

!"#$%&''()

54 Roman Shapiro

(155) Maja g!vali esi tebe shibuka lan’che perishola esi. my tell REL you very before come.PAST REL ‘I have told you to come much earlier’. (Shprintsyn 1968)

Inceptive [INCP] particle la:

(156) Uklainza kakoj godu-la is’o des’a lapaxoza laobodaj. Ukrainian which year-INCP all here logging enterprise work ‘!e Ukrainians — which year was it now? — worked at the logging enterprise’.

(Kukan)

(157) Tam zimli ch’o-ch’o sadi, kukuluza la. there land something-something plant maize INCP ‘He planted something on that land [which had been a tomb], some maize’. (Kukan)

Past tense [PAST] use of verbs with -la borrowed from Russian as ready words:

(158) S’orem’ bali, eta lichila. all.the.time hurt this treated. ‘He was ill all the time, and they treated him’. (Kukan)

(159) Ty perca sadila, eta zel’ona? you peppers planted this green ‘Have you ever planted peppers, the green ones?’ (Kukan)

(160) Lan’she ja pashol! toka lipaxoza desi. before I went only logging.enterprise here ‘I used to go only to the logging enterprise before’. (Kukan)

Past tense [PAST] particle byla:

(161) Balagura konchaj byla. joke end PAST ‘He stopped joking’. (Cherepanov 1853)

(162) Majá sa tyvajá včala χatí bylá. my for your yesterday go PAST ‘I was looking for you yesterday’. (Aleksandrov 1884)

(163) Ja sor!f!taroj goda, ja pashola Perpidz’an byla, ja nachanika tam pashola. I forty-second year, I went Birobidzhan PAST I boss there went. ‘In 1942 I went to Birobidzhan to the boss’. (Kukan)

Negation (see Section 4.3.2):Particle netuA. General negation:

(164) Za moja fal’shivajla nitu. TOPIC my false NEG ‘I am not telling lies’. (Cherepanov 1853)


(165) Za zhenusheki mes’aca posidi netu. TOPIC wife together sit NEG ‘!ey do not sit together with their wives’. (Cherepanov 1853)

(166) Maja igylaj netu, symies’a netu. I play NEG laugh NEG ‘I am not kidding, I am not laughing’. (Shprintsyn archive)

B. Non-existence:

(167) Isho sange41 netu-la. still three.CL NEG INCP ‘!ree [things] are still missing’. (Shprintsyn archive)

(168) Uot !ta ush konchila-la, qasa netu. INTJ this already end.PAST-INCP, now NEG ‘We’ve already run out of it [ginseng], there’s nothing le%’. (Kukan)

Adverb ni, ne:

(169) Za moja tako ne xychi. TOPIC my so NEG want ‘I don’t want to do so’. (Cherepanov 1853)

(170) Xadi ne nada. walk NEG need. ‘You don’t need to go’. (Shprintsyn archive)

(171) Des’a biz rukavisa ni mozhina. here without glove NEG possible ‘You can’t do without gloves here’. (Kukan)

(172) Pobol’shane kurema adali vorota ne zapiraj esa big jacket as.if gate NEG shut REL ‘!e big jacket is not buttoned up, just like an open gate’. (Cherepanov 1853)

(173) Maja taka vidi tyvaja ni bili. my so see your NEG take. ‘I could see you wouldn’t take it’. (Shprintsyn archive)

SyntaxBasic word order, two-role sentences:

(174) Balagan najdi, suxie drova najdi, spichki najdi, kushaj najdi– propadi netu! hut "nd dry "rewood "nd matches "nd eat "nd perish NEG ‘We’ll "nd a hut, we’ll "nd dry "rewood, we’ll "nd matches, we’ll "nd food — we won’t

perish!’ (Arsenyev 1983)

41. From the Chinese 三个 san1ge ‘three’+classi"er.

!"#$%&''()

56 Roman Shapiro

(175) Obmani ponimaj, serdis’ ponimaj, krugom ponimaj! cheat understand, get.angry understand, around understand ‘It [a boar] knows how to cheat, it knows how to get angry, it knows everything!’

(Arsenyev 1983)

(176) Emu mogu krichi, mogu plakat’, mogu tozhe igraj. he can shout can weep can also play. ‘It [water] can shout, it can weep, it can also play’. (Arsenyev 1983)

(177) Ja sor!f!taroj goda, ja pashola Perpidz’an byla, ja nachanika tam pashola. I forty-second year I went Birobidzhan PAST I boss there went ‘In 1942 I went to Birobidzhan to the boss’. (Kukan)

(178) Soi sadili, sinic! sadili, kukuruza sadili dl’a kolxoz!. soya plant.PAST wheat plant.PAST maize plant for collective.farm ‘We planted soya beans, we planted wheat, we planted maize for the collective farm’.

(Bikin)

(179) Uot! sejchas! prodajot us’o !to, dr’an’ !ta. INTJ now sell all this, rubbish this.

Prodajot ketu, razn!, ikru. Nu desi v magazinax prodajot. sell chum.salmon caviar INTJ here in shops sell ‘!ey sell it all, this rubbish. !ey sell chum salmon, caviar and so on. !ey sell it in

the shops here’. (Bikin)

(180) Bol’shoj ne umej ego, xoroshij ne umej. big NEG can he good NEG can ‘She couldn’t sew big [things], she couldn’t sew well’. (Bikin)

(181) Kolxoza rabotaj ego, magazin tam. collective.farm work he, shop there. ‘He worked at the collective farm, at the shop there’. (Bikin)

(182) Uot tam i dzili. Na palatkax dzili. Na jurtax dzili, ide popali. INTJ there just live. On tent live. On yurt live, where occur ‘We lived right there [in wooden houses]. We lived in tents. We lived in yurts [= no-

mad’s tents], everywhere’. (Bikin)

(183) !to moj baba… medved’a bil, shibko mnogo. this my wife bear hit very many ‘!is is my wife… [she] killed bears, a lot of them’. (Shishkov 1989)

Location/direction markers:

(184) Lan’she ja poshola toka lipaxoza des’a before I went only logging.enterprise here ‘I used to come only to the logging enterprise before’. (Kukan)

(185) Jivo s’uda xodi Vostok he here walk Vostok ‘He came to Vostok [village]’. (Bikin)


Subject/object markers:

(186) Tam ivo kitajsa pampusheki kushi, aga there he Chinese bun eat yeah ‘!e Chinese ate buns/baozi there, yeah’. (Kukan)

(187) Kad! ibenk! na kitajsa zimli byla ivo, ja tak! delaj when Japanese on Chinese land was he I so do ‘When the Japanese were on the Chinese land, I did so [moved to Russia]’. (Kukan)

(188) Brata ivo, brata pomirala brother he brother died ‘Brother, brother died’ (Kukan).

(189) Brata syna rebiatis!ka is’o polno ivo. brother son children still plenty he ‘My nephew has plenty of children’. (Kukan)

(190) Desi duva l’udi p!lotnika ivo. here two people carpenter he ‘Two men were carpenters’. (Kukan)

Word order in questions:

(191) Ty videl karchika? you saw photo ‘Did you see the photo?’

(192) Papushk!, ty ch’o delaj? grandma, you what do ‘Grandma, what are you doing?’ (Kukan)

Postpositions:

(193) Ivana doma sepasibo Ivan house thanks ‘!anks to Ivan’s house [or familiy?]’ (Cherepanov 1853)

(194) Za vashe zh’onusheki mes’aca posidi esa? TOPIC your wife together sit REL ‘Do you sit together with your wives?’ (Cherepanov 1853)

(195) Zimli nizu ground below ‘Under the ground’ (Kukan)

(196) Sorok div’ata ivo kampani tam zhivi forty-nine he together there live ‘Since 1949 I lived with her there’. (Kukan)

(197) Nasha ivo ganbandi laobodaj. our he together work ‘I work with him’. (Shprintsyn archive)

!"#$%&''()

58 Roman Shapiro

(198) Karytoshka kamypani kushi. potatoes together eat ‘We’ll eat it with potatoes’. (Shprintsyn archive)

Prepositions:

(199) Na tibi posemoteri, adali ledeneza kushaxu esa. on you look as.if lollipop eat REL ‘Looking at you is like eating a lollipop’. (Cherepanov 1853)

(200) Za evo malen’ki kurema pode-bol’shana kurema ponosi esa. TOPIC he small jacket under-big jacket wear REL ‘!ey wear a small jacket [i.e. a waistcoat] under a big jacket [i.e. a tailcoat]’

(Cherepanov 1853).

Comparatives: (201) Pyluga kareica adinaka plough Korean like ‘plough like a Korean one’ (Shprintsyn archive)

(202) Ivo iseravno Leni, Leni kamypani. he same Lenin, Lenin together. ‘He [Sun Yat-sen] is like Lenin, together with Lenin’. (Shprintsyn archive)

(203) Za nashi zhenusheki adali boga podumaj esa. TOPIC our wife as.if god think REL ‘We think our wives are like goddesses’. (Cherepanov 1853)

China Coast PidginTexts and contexts

Umberto Ansaldo, Stephen Matthews, and Geo& SmithUniversity of Hong Kong

In this paper we revisit some long-standing questions regarding the origins and structure of China Coast Pidgin (CCP), also known as Chinese Pidgin English. We "rst review the historical context of the China Trade which formed the ecol-ogy for the development of CCP. We then review the available sources, focus-ing on newly transcribed data from Chinese-language instructional materials. !ese sources provide fresh evidence for grammatical structure in CCP, and demonstrate strong in$uence from Cantonese as the major substrate language. Comparison with English-language sources shows systematic contrasts which point to likely variation between Anglophone and Sinophone lects, as in the case of wh-questions which show regular wh-fronting in English sources and pervasive use of wh-in-situ in Chinese sources. !is conclusion helps to resolve the debate over the Sinitic features of CCP.

1. Introduction

China Coast Pidgin, also known as Chinese Pidgin English, has been called the ‘mother of all pidgins’ for a number of reasons (Li, Matthews, & Smith 2005: 79).1 In particular, it seems to have given the name ‘pidgin’ to functionally restricted codes that arise in predominantly bilingual communities (Reinecke 1937). China Coast Pidgin (CCP) has also been associated with the expansion of English-lexi"er pidgins in the Paci"c (Siegel 1990; Tryon, Mühlhäusler, & Baker 1996: 485). While typically stereotyped and ridiculed in the popular British imagination and litera-ture, it functioned as a lingua franca in the communities of Western traders found in the ports of the China Coast (and some inland markets). An unusual feature of CCP, particularly important for linguistic scholarship and the study of pidgins, is

1. We term the contact language at issue ‘China Coast Pidgin’ to specify its coastal origin and to avoid the implication that it is a variety of English.

!"#$%&''()

60 Umberto Ansaldo, Stephen Matthews, and Geo& Smith

that it was available in written sources for the purpose of transmission among the Chinese (Shi 1993).

!is paper o&ers a historical discussion of the context of formation of CCP, and a structural description of the language as represented in marginalia in "e Chinese-English Instructor, the largest single source of credible CCP texts, and one rich in grammatical data. Our aim is to add to our knowledge of the evolution of CCP, on which relatively little has been written. In previous scholarship, there has been a tension between substratist analysis (Shi 1991) and claims of English in$uence and universals of pidginization (Baker & Mühlhäusler 1990) in account-ing for the evolution of the pidgin. !is paper presents strong evidence for the latter, if we allow for the existence of substantial variation in CCP. !e paper is structured as follows. Section 2 is a historical introduction to the Western pres-ence in Southern China and an in-depth look at the ecology of the Pearl River Delta, where CCP was born. Section 3 describes the sources for CCP available to date and summarizes previous scholarship on the topic. Section 4 is a grammatical analysis of the Instructor data. Section 5 discusses the "ndings in relation to previ-ous studies on CCP and draws our conclusions.

2. History of CCP

!e history of Macau and the Pearl River Delta constitutes probably the most fundamental chapter in the development of East-West contacts in China and the linguistic ecology that resulted from these encounters. As a consequence of these East-West encounters, at least two new varieties emerged between the seventeenth and the eighteenth centuries: (1) Makista, an Asian-Portuguese vernacular with Sinitic elements that became the idiom of the mixed Portuguese-Asian commu-nity of Macau (see Ansaldo & Matthews 2004; Ansaldo, Matthews, & Smith 2009; Ansaldo 2009) and (2) China Coast Pidgin (CCP), the trade jargon that devel-oped around the Canton (Guangzhou) area during British colonial rule, and later spread to other ports of China and even into the Paci"c.2

2.1 !e British and Portuguese presence in Southern China

!e British arrived in Southern China over a century a%er the Portuguese. Although there were sporadic contacts with the Ming and Southern Ming Dynasties, most of the early interaction was in the period 1644–1684, at which

2. One indication of this is the word kaukau in Hawaiian Creole English, derived from CCP chow-chow ‘food’.

China Coast Pidgin 61

time several problems a&ected the newly installed Qing dynasty. Besides the usu-al border wars against Altaic tribes in the northwest, Chinese authorities were busy dealing with the uprising of the Taiwanese under the ‘pirate’ Koxinga, which among other things contributed to the ban on maritime commerce. Chinese authorities were thus unconcerned with the new traders, and happy to use the Macanese — Macau residents mainly of Portuguese ancestry — as mediators in the initial contacts. Macanese authorities were interested in protecting their mo-nopoly on trade with China and attempted to con"ne and control the activities of the British as much as possible. !e British resented Portuguese mediation and tried hard to trade directly with Chinese authorities. In 1685 an imperial decree determined that Canton would be opened for trade, and in 1699 the British were allowed to establish a ‘factory’ in which to conduct business (Gunn 1996: 27); this is a term derived from Portuguese feitor, an agent dedicated to the trade in goods under royal monopoly, a central "gure in Portuguese mercantile history since the early "%eenth century (Paviot 2005: 24). From this period come the comments of Mundy (1637), who noted the di#culties encountered when trying to use English to communicate with the locals. Other travelers reported the use of a mixture of English and Portuguese in the interaction with local Chinese well into the middle of the eighteenth century (Noble 1762). Obviously Macau played an important role in the early trade between Western merchants and China. Among the sons of the Macanese (or casados), a class of interpreters grew over time, specialized in mediating between the Chinese authorities and Portuguese o#cials. !ey were $uent in written and spoken Chinese, most likely both Cantonese and Mandarin as they were dealing with local authorities, who would require the former, as well as court o#cials from the capital, for whom only the latter would be acceptable (Boxer 1973). !is linguistic expertise positioned them at the center of the East-West trade as interpreters and mediators, a role that they held on to even a%er the establishment of Dutch and British trading houses led to a decline of Portuguese economic power. Not only was a form of vernacular Asian-Portuguese the lin-gua franca for Western traders, it was also the best means of communication with Chinese parties until the late eighteenth century. Moreover, the Portuguese were regarded as the best source of information on Chinese (as well as Japanese) cul-ture and manners, because it was in Macau that the only schools for the study of Chinese language and culture could be found, especially linked to missionary ac-tivity (Martino 2003; Tamburello 1983). As stated in the Chinese Repository (1833, cited in Martino 2003: 21), ‘for over a century from 1517 the only European ships to visit China were Portuguese, and their language became, to some extent, the lin-gua franca of the coast’. Finally, even a%er the opening of Canton, Macau remained a periodic residence of Western traders, in particular each summer and autumn,

!"#$%&''()


when the city would "ll up with traders waiting for their boats to be measured or papers to be issued (Gunn 1996: 28).

It is only from the latter part of the 17th century that British presence started growing in Southern China, as a result of (a) the refusal by Japan to trade with the British, which led to a growing interest in the China trade, (b) the recognition by other Western trading powers of the East India Company, and (c) the resolu-tion of the con$ict between China and Taiwan, which opened up maritime routes once more (Martino 2003). For these reasons, in the "rst phase of British-Chinese relations the role of English would have been irrelevant (Martino 2003: 24); not only was Portuguese mediation central in this phase, but British presence in the region was also severely limited and lacked constant presence, as the British were still very focused on India and North America until the end of the seventeenth century. !ough it is impossible to rule out British-Chinese linguistic contacts in this early phase, it is much more likely that the trade jargon known as China Coast Pidgin evolved in a later period, following the ‘Flint case’ and the development of the ‘Canton System’ discussed in the following section (Ansaldo 2009).

2.2 !e Canton System

Already in the early days of the Canton Trade, commerce was regulated by Chinese authorities and was con"ned to a number of limited ports where Western mer-chants carried out business through local Chinese merchants, among whom cor-ruption abounded. In an attempt to liberalize and improve commerce, James Flint, who resided in China between 1736 and 1762, traveled to Tianjin to address the emperor on the question of corruption of the local Cantonese authorities (Martino 2003: 32; Soothill 1925). He obtained help in translating a petition on corruption in the Canton Trade and presented it to the emperor. !e consequences of this interference were drastic: Flint was imprisoned in Macau for three years and then expelled from China; all ports except Canton were closed to foreigners; the trans-lator of the petition was beheaded; and it was forbidden to teach foreigners the Chinese language, on penalty of death. !is also led to the formalization of the ‘Canton System’ and resulted in the fact that, until 1842, all power was virtually in Chinese hands, with the movements of Westerners heavily restricted, contact with the local population was prohibited except through o#cially appointed mediators, and an o#cial position of inferiority was accepted by Western traders for the sake of economic pro"t. It should be acknowledged that many missionaries did manage an impressive command of the language, and presumably the banning of teaching could be circumvented where necessary. Nevertheless, with such o#cial measures in place, the British generally never gained any knowledge of the Chinese lan-guage, nor did they set up institutions to study Chinese culture.


!e ‘Canton System’, which actually evolved during the "rst sixty years of the 18th century, is well-known to historians, thanks to the wealth of records le% be-hind by the European trading companies as well as Chinese o#cial documents (Van Dyke 2005). !e Canton System developed in order to gain better control of the activities of Westerners and local o#cials in the Pearl River Delta. While it was di#cult and potentially expensive for the Chinese to exert total control over the comings and goings of boats in the open sea o& Macau and its islands, as this would have required extensive patrolling, forcing boats to sail up the Pearl River into Canton provided the authorities with an ideal passage which was easy to con-trol (Van Dyke 2005). It was organized along typical Chinese lines (Fairbank & Goldman 1998: 195): Chinese merchant families were appointed by the govern-ment to act as brokers and superintendents of the foreign traders. Each foreign ship was the responsibility of one Chinese family-"rm. !ese ‘security’ merchants formed a guild, the Cohong, at the command of a customs superintendent, and the Hoppo, who answered directly to the Emperor. !e Cohong and the Hoppo as-sisted, controlled, and taxed the foreign cargoes. !is proceeded in three steps: (1) foreign boats were required to call at Macau and were then escorted up the river to Canton; (2) at the harbor of Whampoa (Huangpu), 20 km from Canton, the boats stopped and the cargo was unloaded onto smaller vessels; (3) from Whampoa, the chief trader of the foreign "rm (the taipan or supercargo) and the cargo moved to Canton, while the ship and crew were le% at Whampoa.

During the 18th century there were no more than 20 supercargoes, and very few of them visited the foreign quarter at Canton, typically only once every three years due to prohibition or time spent at sea (Li et al. 2005; Shi 1986). Until the end of the eighteenth century, ships’ crews had to remain in the port of Whampoa for three months during winter; there could be up to 2,000 British sailors at a time as well as other foreign crew, and we know that interaction with the locals, in terms of prostitution, street "ghts, etc. was common (Bolton 2003: 157; Martino 2003: 51). One of the most notorious sentences in early CCP, in which a rare mix-ture of Portuguese and English elements is found, certainly comes from this type of environment: carei glandi hola pickenini hola?, ‘you want the big whore or the small whore?’ (Noble 1762: 240). In the mid-19th century, it was reported that in Whampoa the Chinese could speak better English than in Canton (Nicol 1822; Williams 1836). !is implies that, notwithstanding the tight o#cial control, an un-derworld of material and human smuggling allowed contacts between Westerners and Chinese to take place.

!"#$%&''()


2.3 !e nature of Chinese-Western contacts

In order to appreciate the history of East-West encounters in general, it is im-portant to realize that Western arrivals in Southern China were not always seen as intrepid adventurers, lucrative business-partners, or bearers of higher knowl-edge. China’s relations to non-Chinese, in particular towards populations of the Western borders (and beyond) has always been clear: outsiders were barbarians, a potentially polluting element for the superior Chinese culture, to be kept at bay at all costs. Chinese authorities had been intent on limiting and controlling interac-tions between Chinese and outsiders since the 15th century. !e same treatment was applied to the Portuguese, the Dutch, and especially the British, as revealed in historical accounts of East-West relations between the seventeenth and eighteenth centuries.

In his historical analysis of East-West contacts in the nineteenth century, Fairbank (1953) makes it clear that, from the Chinese point of view, Westerners were inferior in terms of culture:

!e decadent, part half-caste community of Macau remained walled o& on its peninsula; the thirteen Factories at Canton were outside the city walls, from which foreigners were excluded. All contact with foreign merchants was mediated through a special class of compradors, linguists, shro&s and Chinese merchants, as well as through a special language. (Fairbank 1953: 13, our emphasis)

Western men were depicted in Chinese sources as ‘…violent and tyrannical and skilled in the use of weapons… !ey wear disheveled hair which hangs over their eyebrows’ (Chen 1939: 347). Western missionaries and adventurers since the 15th century had described China as a rich and impressive country, where mate-rial wealth, architectural wonders, and advanced scienti"c knowledge were to be found (Boxer 1953). !e feeling of superiority that the Chinese displayed towards Europeans was well captured by !omas Taylor Meadows, a British interpreter and student of China, who in 1852 wrote:

!e Chinese do habitually call and consider Europeans ‘barbarians’, meaning by that term ‘peoples in a rude, uncivilized state, morally and intellectually unculti-vated’… !ose Chinese who have had direct opportunities of learning something of our customs and culture — they may amount, taking all Five Ports, to some "ve or six thousand out of the three hundred and sixty millions — mostly consider us beneath their nation in moral and intellectual cultivation. As to those who have had no such opportunities, I do not recollect conversing with one… !ey are always surprised, not to say astonished, to learn that we have surnames, and understand the family distinctions of father, brother, wife, sister, etc.; in short, that we live otherwise than as a herd of cattle.


!e only Western group that had managed to break through the iron curtain that the Chinese had put up against foreigners throughout their history were mission-aries, in particular the Jesuits, but even they were eventually dismissed. !is oc-curred a%er they lost the ‘controversy over rites’, a public debate between scholars of di&erent religious traditions held by the imperial court in 1725, and they be-came virtually powerless a%er the rejection by the ‘Son of Heaven’ of the Pope’s ecumenical claims (Fairbank 1953: 14).

As a result of such attitudes, Chinese-British relations in the Canton System were in fact not symmetrical. In contrast to what has been suggested in Stoller (1979) and followed up in Tryon et al. (1996), British-Chinese communication in this phase may not constitute an example of ‘high-high’ exchange. Until the "rst Opium War of 1842, the British (and other Western traders) were most certainly in a lower cultural position, due to the general contempt that Chinese had of foreign-ers, and the superior status of the Portuguese in linguistic and cultural matters. Since the British were in a politically and economically disadvantaged position, limited by the Canton System which was designed to favor Chinese authorities, we can hardly speak of mutually bene"cial exchanges: British and other foreign merchants needed China much more than China needed them, since the revenues from the trade with the West were marginal to the Chinese when compared to the revenues of their Asian trade (Fairbank 1953). In fact, it was exactly because of this overall position of inferiority that the British challenged Chinese authority on such a shady issue as the Opium Trade, and used it as an excuse to "ght the "rst of a series of wars leading to the ‘unequal treaties’ that forced China to negotiate terms of trade favorable to Western powers. Considering the overall inferior posi-tion of the British in this phase, if a pidgin English had already started develop-ing, it makes sense to revisit Hall’s (1944) suggestion that the Europeans played a signi"cant role in ‘pidginizing’ their own language, desperate as they were to engage in trade with a superior power which held them in contempt and was only marginally interested in their trade (Chinese trade with the Philippines, Southeast Asia, Central Asia, and Japan being far more pro"table). In addition, considering that it was not only the British but also other Western traders who used English among each other as well as in their contacts with the local populations, in an at-tempt to escape Portuguese control, the theory of Westerners pidginizing English in their interactions with Chinese makes sense, and lends further credibility to the hypothesis of European agency in the development of a pidgin variety (Hall 1944).

Clearly the Canton Trade would have o&ered opportunities for some degree of language contact between Asian-Portuguese, English, Cantonese, and perhaps o#cial Chinese (guanhua or Mandarin). Interaction between pilots and the crew would have o&ered a setting for Makista and English to come into contact. !e com-pradors, despite the limitations imposed on the factories, were a point of contact

!"#$%&''()


between English and Cantonese; moreover, they could provide, in uno#cial ways, further opportunities for local Chinese to come into contact with Western traders. Interpreters (‘linguists’) clearly played a signi"cant communicative role, potentially bringing Chinese, Makista, and English into further contact, though they were few in number and acted in very limited contexts (Van Dyke 2005: 81).

We should however not think of the Canton Trade as the only locus of contact: the Opium Trade, which involved a great deal of smuggling, put Western traders and local interpreters, buyers, and dealers into contact in various coastal areas of China (Fairbank 1953). Other aspects of illicit commerce would have had the same e&ect; Chinese parties also engaged in illegal trade as this would result in much higher pro"ts for themselves. Corruption was indeed so high that eventually it brought the Canton System to a halt (Fairbank & Goldman 1998).

!e complexity of interactions in the early phase of East-West encounters em-phasizes the importance of historically-nuanced approaches to contact ecologies, as advocated among others by Faraclas et al. (2007) and their notion of ‘co-habi-tation societies’, in which micro-ecologies of informal contacts are given serious consideration. While contact between English and Cantonese must have taken place, it is clear from the above that such contacts were neither frequent nor pro-longed in time until the late eighteenth century. !e contacts between English and Makista, it seems, must have been just as super"cial, short-lived, and infrequent (for this see the discussion in Ansaldo et al. forthcoming).

2.4 Development and decay of CCP

Between the 18th and 19th centuries, several political and economic changes took place that resulted in a great in$ux of trade. !e First Opium War (1839–42), fought in order to force Chinese authorities to release monopoly on commerce and to liberalize the Opium Trade under British control, saw the British emerging as victors who forced China to open a total of "ve ‘Treaty Ports’ (Canton, Amoy, Fuchow, Shanghai, and Ningpo), where the British were free to trade with any Chinese party. It also resulted in Hong Kong becoming part of the British Empire. With further concessions resulting from the second Opium war (1856–1860), Western trade expanded along the China coast as well as inland.

!e role and function of the East-India Company had already been increasing since the late eighteenth century, due to an expansion of trade and further recogni-tion of its key role by the international community. !is was one of the three main contexts of English usage in Southern China at the beginning of the 19th century identi"ed by Bolton (2003), the other two being the activities of country traders and Americans. As a consequence of such expansion, the community of the factories based in the foreign quarter of Canton started growing and social relations increased


in complexity. Martino (2003: 62) captures the contacts between Westerners and Chinese in the Canton area as seen in Table 1 (see also Bolton 2003: 157).

In 1836 the Anglophone community in Canton comprised approximately 200 individuals, the majority of them males — Western women tended to reside in Macau3 — with a further 100 individuals, over half of them of Indian/Persian ori-gin, and the rest Western European, predominantly Portuguese, living in the for-eign quarter (Morse 1926). In 1827 the "rst English periodical appeared, which is o%en taken as an indication of an expansion of the foreign community. In 1836–7 two articles published by Wells report that CCP booklets, as discussed in the next section, were in circulation among the Chinese population. !e existence of such manuals testi"es to the importance of CCP, as well as to the high level of education and the importance of instruction in the Chinese community. In the late 18th cen-tury, as a consequence of the Opium Wars and certainly also with the help of such booklets, CCP spread to other parts of China (e.g. Shanghai). CCP was reportedly used also among Chinese of di&erent linguistic backgrounds and of course within the non-English speaking Western community.

It is clear that despite o#cial controls there was interaction between Western and local Chinese in and around the foreign quarter of Canton in this phase (Hunter 1885). For example, within the factories, merchant-servant interaction would have brought Chinese and English into contact. As noted in Selby & Selby (1995), inside the trading houses the relationship between merchants and their employees was quite varied; there were not only cooks and cleaners but also what might be seen as apprentices, assistants, o%en related to the Chinese compradors, who were in the service of the foreigners for the purpose of learning and acquir-ing wealth through their own trade. !ese were o%en of higher social extraction and well educated. It is indeed to one of these key "gures that the compilation of a CCP glossary is assigned. Outside the factories, interaction with shopkeep-ers and coolies would have taken place. With commerce intensifying, in the port

3. !e British had a limited presence in Macau which lasted until the British took over control of Hong Kong in 1842 (Coates 1966; Tryon, Mühlhäusler, & Baker 1996).

Table 1. Contacts in the Canton region (adapted from Martino 2003: 62)Factories WhampoaFormal relations

Foreigners Informal relations

Formal relations

Foreigners Informal relations

TradersCompradorsInterpreters

SupercargoInterpretersMissionaries

ServantsCompradorsShopkeepersCoolies

Compradors Sailors Local sailorsProstitutesOther locals

!"#$%&''()


of Whampoa informal relations between crews and locals must have sustained a multilingual ecology.

CCP started declining in the twentieth century. !is is usually attributed to the impact of English-medium instruction becoming available in missionary schools (Bolton 2003: 191). However, another factor should be taken into account, namely the dissolution of CCP’s ecology. First, there was the decline of the Canton trade around 1830, as a result of the East India Company closing their activities there (Van Dyke 2005: 175); CCP nonetheless continued to be spoken in other trading enclaves, including Hong Kong and Shanghai. Secondly, World War II and the Japanese occupation caused a dramatic change in the ecology of Chinese coastal areas, in particular Hong Kong (Mühlhäusler & Baker 1996: 518). While in the twentieth century, CCP could be found in Hong Kong, Shanghai, and Macau spo-ken amongst shopkeepers, domestic servants, and sailors (Mühlhäusler & Baker 1996: 517), a%er the Japanese occupation a new wave of expatriates replaced the community that had been displaced during World War II. Even if knowledge of CCP had survived among shopkeepers and domestic servants, the new expatriates had no knowledge of CCP and this would have led to its decline. A%er World War II, Hong Kong rapidly moved towards a modern and cosmopolitan society. !e complex, informal ecologies of the Canton system, the illegal trades run by mer-chants of varying social extraction and with di&erent educational backgrounds, the interaction of di&erent European vernaculars and Chinese languages, and the mediation of more or less competent middlemen all came to an end then. From the 1960s, the population of Hong Kong increasingly shi%ed towards an Anglo-Chinese bilingual (and bicultural) identity. !e new expatriate contingent spoke various English varieties, all more or less in standardized form, which also served as the target of teaching in local schools. !e ecology of CCP had therefore de facto dissolved, and with it the raison d’etre of CCP.

3. CCP sources and previous scholarship

An annotated corpus of all currently known attestations of CCP from English language sources was assembled by Philip Baker in the 1980s, and although this remains unpublished (see Baker 1987: 164 for details), it has been made available to other researchers, and this has been our main point of reference for sources in English. !is corpus of around 15,000 words contains data varying in qual-ity, as it was o%en recorded incidentally, for example when illustrating humorous stories of encounters between Chinese and trading partners or travelers. During the expansion of CCP use, a series of booklets appeared, written by Chinese in Chinese characters for the purpose of teaching and learning CCP. In contrast to


the attitude of many English speakers, the Chinese do not appear to have regard-ed the language as a source of humor, and these publications suggest the serious purpose of achieving upward mobility through knowledge of the linguistic me-dium required for trade. Among the many sources, two are known to us today: (1) Hùhng mòuh tùng yuhng fàan wá (紅毛通用番話) ‘!e Language of the Redhaired Foreigners’ (henceforth Redhaired Glossary) (anonymous c. 1835) published in Canton around the fourth decade of the 19th century, and (2) Yìng yúh jaahp chyùhn (英語集全), the much less widely available six volume work Chinese and English Instructor (henceforth Instructor) (Tong 1862), written by Tong King-Sing around 1862 (see also Leland 1892; Selby & Selby 1995; Williams 1836). CCP was reportedly used not only in East-West encounters and within Western circles but also among Chinese of di&erent linguistic backgrounds (Whinnom 1971: 104). We must therefore recognize a fourth group of CCP users next to the three mentioned above for the nineteenth century: Chinese traders, merchants, and servants.

3.1 !e Redhaired Glossary

Two minimally di&erent versions of a booklet with this title can be found in the British Library. 紅毛通用番話 (hùhng mòuh tùng yuhng fàan wá) literally means ‘the language of the redhaired foreigners’, i.e. Europeans. An earlier version had a slightly less complimentary title with 鬼 gwái ‘ghost, devil’ in place of 番 fàan ‘foreign’. It appears that several versions of the booklet appeared, and that similar booklets giving guidance in speaking basic Macau Portuguese were produced as early as the 1750s (Baker 1989: 3). !e British Library booklet, probably dating from around 1835–50 is the source of the examples given here. It has 16 pages of words or short phrases dealing with numerals, occupational vocabulary, etc. A sample page is shown in Figure 1.

For each entry, the term is given in Chinese, and beneath it and slightly o&set to the right, the CCP pronunciation is given as closely as possible in Chinese char-acters. A section from page 5 is reproduced in Figure 2.

Two entries from these pages are presented here, with the translation of the Chinese, the Yale transcription of the pronunciation, and the approximate (pid-gin) English equivalent added in parentheses.

水老手 (séui sáu = sailor) 婆 (lóuh pòh = wife)

些威利父 (wài fuh i.e. wife)文 (sè leih màhn = sailorman)

!"#$%&''()


!e booklet contained a total of nearly 400 entries, all single words or formulaic expressions. !us, while this is a valuable source with respect to the lexicon and phonology, it says little about the grammar of the language.

A number of analyses of the Redhaired Glossary have appeared. Shi (1993) gives some useful phonological correspondences, while a more thorough analysis of the content was made in Baker (1989) with a detailed translation and suggestions for the etymology of most of the entries as well as further discussion of phonologi-cal equivalents. !e Yale transcription of the pronouncing characters and target

Figure 1. Page from the Redhaired Glossary

Figure 2. Sample entries from the Redhaired glossary


English or pidgin forms for the entire 372 entries are listed as an appendix to Bolton (2003) who produces some further suggestions regarding the source of certain items in a set of footnotes, including a few possible derivations from Swedish. As noted by Shi (1993), learning a pidgin language through written characters is highly unusual in the context of how pidgin and creole languages are normally acquired.

3.2 !e Instructor

A much more substantial work was produced, probably in 1862 (Selby & Selby 1995: 123) by Tong King-Sing (唐景星), also confusingly known as Tong Ting-ku or Tang Tingshu. Tong appears to have been an accomplished English speaker with a linguistic insight far ahead of his time. He produced a monumental six-volume work, the Chinese-English Instructor (英語集全 yìng yúh jaahp chyùhn), herea%er "e Instructor, which attempted to make Standard English comprehensi-ble to Chinese speakers. Each of the six volumes deals with a di&erent subject area. In addition to single words and phrases, more extended sentences and chunks of dialogue are featured in volumes 4 and 6. In addition to its undoubted linguistic value, the snatches of dialogue also serve as an interesting insight into the ev-eryday life of traders in both camps during the period. !e pronunciation of the English forms is indicated not only by the nearest equivalent Chinese character, but also by a series of diacritics, which appear to use conventions developed in describing minority languages within China (Selby & Selby 1995: 125). A sample page is shown in Figure 3.

Each entry consists of four parts. On the bottom right hand side is the term in English, while on the bottom le% is an indication of the pronunciation in English as far as is possible using Chinese characters. !e constraints of Cantonese pho-nology do present considerable di#culties, but Tong had devised some special symbols, such as a small triangle to indicate inter-dental fricatives, which are not present the Cantonese phoneme inventory, seen in Figure 4.

In the top le% hand area of each column is the term in Chinese, while at the top right is a Romanized equivalent to indicate the pronunciation in Cantonese, as in Figure 5.

However, the item of greatest interest to scholars of CCP is the collection of hand-written marginalia, apparently representing the featured item in pidgin, or as Tong terms it 廣東番話 gwóng dòng fàan wá ‘Canton foreign language’. Two of the items on the right hand page above are shown in close-up in Figure 6.

One of these entries can be analyzed as follows:

!"#$%&''()


How many dollars is that咁申得幾多員呢(gam sàn dàk géi dò yùhn nè)口乜治打鎼(háu màt jih dá làh)CCP: how muchee dollar

A full account of "e Instructor has yet to be written, and there is a great deal of valuable information awaiting analysis. !e Romanized Cantonese, for example, is a signi"cant resource for considering phonological change in Cantonese over the last 150 years. With regard to the CCP marginalia, a start was made by Li et al. (2005), who assembled all the entries for a preliminary analysis (see also Ansaldo forthcoming; Ansaldo et al. 2009). A total of just over 1,000 CCP entries were re-corded. !is represents a corpus of some 5,000 words, which is the largest reliable source of CCP texts in existence.4 For each expression gleaned from the margins, four pieces of information were extracted:

4. Leland (1892) is comparable in size but notoriously unreliable, not least because the author does not seem to have traveled as far as China (Selby & Selby 1995: 123).

Figure 3. A page from the Chinese-English Instructor


Figure 4. Representation of standard English in the Instructor

Figure 5. Chinese and Romanized Cantonese pronunciation from the Instructor

!"#$%&''()


1. !e standard English text, exactly as it is presented in the original book (top le%);

2. !e Chinese characters used to represent the equivalent phrase in pidgin (top right): again this is presented as in the original as far as typography allows;

3. A transcription of these characters as read in Cantonese, using the Yale ro-manization (bottom right);

4. A representation of the pidgin in English orthography as used in English-language sources (bottom le%).

A sample entry is given in Table 2 as an illustration.

Table 2. Organization of the CCP corpus (from Li et al. 2005)p. 32 Law Suit (官訟)It is a bad thing to go to law 哥羅必剪威黎必

go law pidgin velly bad gò lòh bìt jín wài làih bìt

While this represents a very valuable source of data, many questions remain. It is not clear whether Tong or someone else added the marginalia, and, crucially, how representative it is of the language spoken at the time is not known. It is also un-clear how closely the Chinese characters are intended to represent the actual pro-nunciation of items, as opposed to the best estimate using the Cantonese sounds available. Be that as it may, if investigating CCP is like studying dinosaurs (Selby & Selby 1995: 113), decoding the Instructor is tantamount to unearthing a complete skeleton:

…what a goldmine this source is: we have extensive recordings of pidgin dia-logues set down by a talented linguist, together with their colloquial English and Cantonese equivalents. (Selby & Selby 1995: 125)

the written fāanwáa texts in the Instructor are probably the largest single source for Chinese pidgin English available to any Pidgin and Creole scholar. (Bolton 2003: 176)

Figure 6. CCP marginalia from the Instructor


Whereas the better known Redhaired Glossary is largely a list of words, the Chinese-English Instructor provides extended dialogues on themes re$ecting the mercantile functions of the language: law suits, selling tea, chartering ships, etc. Consequently, it provides rich insights into the grammar of CCP as spoken by a $uent Chinese bilingual.

3.3 Previous views on CCP

Typically two somewhat divergent positions can be identi"ed in relation to the origin of CCP grammar: (i) a substantial number of CCP features can be traced back to Chinese, especially Cantonese grammar (Bisang 1985; Selby & Selby 1995; Shi 1986, 1991) and (ii) CCP shows little Chinese in$uence, but ‘universals of pid-ginization’ can be detected (Baker & Mühlhäusler 1990; Tryon et al. 1996). In fact, both Chinese and English in$uence can be found, due in part to variation between attested samples of CCP, especially between those in the booklets by Chinese au-thors, and Anglophone CCP, as represented in the Western literature. In addition, it has been noted that if pidginization involves loss of morphology, then it be-comes indistinguishable from the e&ect of feature transfer of isolating typology in the contact situation, in our case in$uence of Chinese grammar (Ansaldo 2007). A related issue is that of agency in the genesis of CCP. While Hall (1944) claims that CCP was the result of foreigner-talk strategies, Baker & Mühlhäusler (1990) sug-gest that it was the product of a ‘high-high’ encounter and was stable in its ‘expand-ed’ version. !is claim may be called into question based on two observations: (1) Western traders were not generally regarded as representatives of ‘high’ culture in the eyes of the Chinese and mostly held in contempt (Chen 1939; Fairbank 1953); and (2) the di&erent sources reveal the existence of signi"cant variation within CCP (Bolton 2003: 161; Shi 1991).

4. Grammar

!e ‘Chinese’ character of CCP has been noted in previous work. In the Instructor, for example, ‘Much of the syntactic structure of Pidgin follows that of Cantonese’ (Selby & Selby 1995: 128). However, some of the more speci"c claims which have been put forward have proved to be controversial. Shi (1986, 1991) attributed sev-eral features of CCP to structural in$uence from Cantonese, which is assumed to be the mother tongue of most non-European users of the pidgin. !e features discussed by Shi include topic-comment structure, the use of piece(e) as a classi-"er (CL), the interrogative pronoun what fashion meaning ‘how?’ (Cantonese dím yéuhng), compounds with ‘man’ as in tailorman and the distribution of belong as

!"#$%&''()


copula, which (as in Chinese) is not used before predicative adjectives. Baker & Mühlhäusler (1990) call a number of these features into question, in some cases based on the occurrence of parallel features in other pidgins: for example, Tok Pisin has compounds with -man which are comparable to CCP tailorman, etc. !ey nevertheless acknowledge that some expressions appear to be calqued on Cantonese; in particular, the use of piece(e) as a classi"er between the numeral and the noun is distinctively Chinese.

4.1 Words and sounds

CCP words are derived mostly from English, with some Portuguese, Malay, Hindi, and as would be expected, Cantonese in$uences. Some examples given by Martino (2003) include: pidgin < business (E.); catchee < catch ‘fetch’ (E.); two muchy < too much ‘extremely’ (E.); Joss < Deus (P. ‘God’); sabbee < saber (P. ‘to know’); chop < chapa (M. ‘chop’); well-known Cantonese words are: taipan (C. ‘supercargo’, lit. ‘big class’) and fankuei (C. ‘Westerner’, lit. ‘foreign devil’). Words of clear Portuguese origin are less common than one would expect considering the potential in$u-ence that Macao could have had in the early days of the Canton Trade: Bolton (2003: 286) identi"es 20–22 Portuguese words out of 372 in the Redhaired Glossary.

It is di#cult to present a de"nitive account of the phonology of CCP for a number of reasons. In the English sources, sounds are likely to have been heav-ily Anglicized, while in the Chinese sources there are a number of other prob-lems: (i) di&erent dialectal readings of the characters yield di&erent sounds, and in some cases it is clear that a character was not used in the Cantonese but rather in the Mandarin reading (Li et al. 2005: 83; Shi 1991), and (ii) the transcription system makes it di#cult to know what the actual pronunciation of words was, as the Chinese characters in the texts may have served to provide an approxima-tion of the sound. !e actual local pronunciation of the characters may well have become the target by users of the booklets, though this is di#cult to establish (Shi 1993: 461). One of the most robust features was replacement of [r] by [l]; the former does not exist in Cantonese and is thus replaced by the only non-nasal sonorant of Cantonese (Shi 1991: 14): rice > lice. Another widespread feature is de-letion of "nal consonants, again a clear in$uence of Cantonese phonology where syllable "nal consonants are restricted to nasals and unreleased stops. Baker & Mühlhäusler (1990) "nd no signi"cant in$uence of Cantonese in the vowel system but heavy restructuring of syllable structure due to in$uence of Cantonese so that CCV > CVCV, e.g.: stop > sitap (Bolton, 2003: 162).5

5. Considering the methodological di#culties in analyzing Romanization, as well as the inher-ent variability to be expected in pidgin phonology, we do not attempt to discuss phonological


4.2 Nouns and NPs

!e structure of the NP reveals variation between Sinitic and non-Sinitic con-stituent order, as well as typical Sinitic features such as the use of the classi"er piecee. !e patterns [NUM-CL-N] and [DEM-CL-N] are typically Sinitic, as in Cantonese:

(1) Yāt go yàhn Nī go yàhn one CL man DEM CL man ‘One man’ ‘!is man’

!is pattern is very frequent in CCP:

(2) You wantchee catchee one piecee lawyer ‘You will have to engage a lawyer’(Instructor IV.32) 6

Apart from piecee, one other item, chop, appears to be used as a classi"er in the [DEM-CL-N] position in the Instructor:

(3) !isee chop tea what name ‘What is the name of this tea?’ (Instructor VI. 15)

!e fact that a single classi"er accounts for the vast majority of cases does not nec-essarily imply ‘reduction’ of the classi"er system of Cantonese. It is true that many more classi"ers are technically available in the language, but in discourse there is an overwhelmingly frequent one of generic use (go in Cantonese, ge in Mandarin; see Erbaugh & Yang 2006 for other Sinitic languages) while the others are more specialized. CCP takes this tendency a step further by reducing the system to vir-tually one classi"er which is not obligatory (see (19) below7 where no classi"er appears in the context [DEM-(CL)-N]).

!e pronouns of CCP have attracted some attention in the literature. In par-ticular, besides what appear to be regular English forms, Baker & Mühlhäusler (1990: 104) and Tryon et al. (1996: 488) note that in a "rst phase three "rst person

characteristics of CCP here.

6. References to the Instructor refer to the volume and page in the original Chinese text, which are also given in Li et al’.s (2005) transcription. !e example sentences are presented as tran-scribed in Li et al. (2005), with some minor modi"cations for ease of reading. !e translations are those given in the original text; more precisely, the standard English translations are the original English examples for which pidgin equivalents are provided in the margin of the text.

7. An extensive discussion of adpositions in CCP can be found in Li, M. 2011a. Chinese pid-gin English and the origins of pidgin grammar. Doctoral dissertation. !e University of Hong Kong. For a speci"c treatment of long see Li, M 2001b. Origins of a Preposition: Chinese Pidgin English long and its Implications for Pidgin Grammar. Journal of Language Contact IV: 269–294.

!"#$%&''()


singular forms were found, I, my, and me. All three of them could be found in subject position, while my and me were used in object position. Eventually, only my emerged in both functions, and this could be taken as evidence of stabilization in the grammatical system of CCP. Smith (2008) shows that while second and third person forms were almost invariant in the Instructor, there was considerable variation in "rst person forms, with some di&erentiation of subject and object forms. !e emergence of ‘my’ as the canonical "rst person form is somewhat sur-prising considering the almost universal adoption of ‘mi’ in pidgins and creoles worldwide. Baker & Mühlhäusler (1990) note that this took place rather rapidly only a%er the 1830s, before which ‘I’ and ‘me’ were normally encountered. !ey hypothesize that the change was brought about by the introduction of written in-structional materials and while the timing strongly suggests this, hard evidence is di#cult to pin down. It is possible that the character 米 which is pronounced máih in Cantonese was entered in the booklets by someone with the Mandarin pronun-ciation [mi] in mind. Another confusing factor is the super"cial similarity in ap-pearance between the character 米 and 未 (mi in nineteenth-century Cantonese), which could have been confused during the printing process.

Compounding is frequently found in CCP, particularly in reference to people and functions. A compound with the form man (‘man’, calqued on Cantonese yàhn) is found: ba ba man, ‘barber’; se lei man, ‘sailor’; guk man, ‘cook’ (from Martino 2003: 86). Other compounds of the type Modi"er-N can also occur, e.g.: Joss pidgin, ‘religion’ (lit. God business); Joss house, ‘temple’. Another calque of a Cantonese word is found in the term ‘fashion’, rendered as fasi (Cantonese yéuhng, ‘manner’, as in dím yéuhng, ‘what manner’ or ‘how’, (Shi 1986), e.g.: wat fa si, ‘what fashion or what way’; niu fa si, ‘new fashion or new way’ (Martino 2003: 87). In the Instructor, how fashion (5) is used in a way parallel to dím yéuhng (5) in Cantonese:

(4) My savvy how fashion do ‘I will know how to act’ (Instructor IV.33)

(5) Cantonese: Ngóh jī dím yéuhng jouh I know how fashion do I know how to do it

4.3 Copulas, zeros, and existence

Baker & Mühlhäusler (1990) report at least two di&erent types of copula: ‘have’, as in (6), usually realized as hab as in (7), this can also be used as a possessive verb as well as an aspectual marker); and belong (13–14). According to Baker &


Mühlhäusler (1990: 103), belong took over as a copula while the aspectual usage of hab increased. Eventually, hab and habgot were used in possessive constructions.

(6) Chinese man very great rogue truly, but have fashion, no can help ‘Chinese men are real rogues but that’s how it is, can’t help it’ (Anonymous

1748, in Baker & Mühlhäusler 1990: 103).

(7) My hap go court one time ‘I have been to court once’ (Instructor IV.4)

Note that copula constructions are relatively rare, as more o%en than not CCP is zero-copula, like Sinitic languages in general, as in (8) (this was already noted in Hall 1944):

(8) Englishman very good talkee; all heart bad, — no talkee true — too much a proudy (Selby & Selby 1995: 138)

!e verb got is likewise used in phrases such as (9), and very similar to the Cantonese verb yáuh (10), which can indicate location, possession and existence (Matthews & Yip 1994):

(9) You got how muchee piecee children ‘How many children have you?’ (Instructor IV.55)

(10) Léih yáuh géi dō go saimānjái? you have how many CL children ‘How many children do you have?’ (Cantonese)

In example (6) above, the second phrase but have fashion, no can help is di#cult to interpret, because fashion can mean many things in CCP and is di#cult to trans-late. For example, in the following: so fashion you buy some beefoo, ‘well, you better buy some beef ’ (Instructor VI.26), the item ‘fashion’ clearly has no literal meaning, as is o%en the case in CCP, but is calqued on Cantonese gám yéuhng ‘so, in that case’. !erefore the phrase but have fashion in the example above could just mean ‘well’, as in Cantonese haih gám yéuhng ‘that’s the way it is’. In the Instructor, most occurrences of hap, many of which occur together with got, follow the Sinitic ex-istential pattern ‘there is’, and are thus not clear cut copulas. !e same was found already in Hall’s data (1944) and strongly suggests that CCP is predominantly zero copula. As far as belong is concerned, it is worth noting that in most occurrences in the Instructor, it is used in its lexical function, ‘to belong to’, as in (11) and (12):

(11) !ese belong you? ‘Is this yours?’ (Instructor IV.53)

!"#$%&''()


(12) !e tea belong "rst crop ‘!is is "rst crop tea’ (Instructor VI.14)

It is unclear to what extent belong really represents a copula or whether a sentence with belong can be read as attributing a relation between two phrases (see Shi 1991: 24) as in the following examples:

(13) You belong honest man ‘you are an honest man’ (Selby & Selby 1995: 136)

(14) !is belong my plum cashee ‘that is what I paid for it’ (Instructor VI.3)

In terms of negation, no (sometimes realized as lo) or no got are used in CCP:

(15) Missy ___ no got houso ‘Miss__ is not home’ (Instructor VI.38)

(16) No got suchee thing ‘there is no such thing’ (Instructor IV.51)

!e negative existential expression no got is typical of Sinitic (Cantonese móuh, Mandarin meiyou). Another reason to be cautious about the status of the copula in CCP is the fact that adjectival and adverbial phrases behave like verbs when used as predicates, a feature that usually correlates typologically with zero-copula, and which is typical of Sinitic (e.g. Cantonese léih hóu faai, [you very fast], ‘you are (very) fast’). !is property is attested frequently in the Instructor:

(17) Court expensee too muchee; the court fees are very heavy (Instructor IV.32)

(18) My too muchee trouble I was very much bothered (Instructor IV.32)

(19) !isee wine glassee no clean this wine glass is not clean. (Instructor VI.47)

4.4 Placement of adverbs and prepositional phrases

Another area showing clear in$uence of Chinese syntax is adverbial modi"cation. Selby & Selby (1995: 128) claim that time and place adverbs follow English syntax, and indeed we "nd such examples as in:

(20) My talkee you tomorrow ‘I will let you know tomorrow’ (Instructor VI.9)


!e place of tomorrow here is not legitimate in Chinese. However, the Instructor also o&ers many examples of sentential adverbs appearing between the subject and verb (21–22). !is is illegitimate in English but in line with Cantonese syntax, as in Ngóhdeih tīngyaht būn [we tomorrow move] ‘we move tomorrow’ (see Matthews & Yip 1994: 187):

(21) We tomorrow makee move ‘We move tomorrow’ (Instructor IV.49)

(22) He every day tipsy ‘he gets drunk every day’ (Instructor IV.55)

From a typological point of view, one of the clearest Sinitic features is the preverbal PP as can be seen in the patterns (23–25) with long below:

(23) My no long you buy anymore ‘I won’t buy anymore from you’ (Instructor VI.26)

(24) You can long my catchee one piecee good boy ‘You can get a good boy for me’ (Instructor VI.51)

(25) My long you takee alla ‘I will buy the whole from you’ (Instructor VI.8)

!e preposition long, derived from English ‘along (with)’, appears in preverbal po-sition like its Cantonese counterpart tùhng ‘with’, and covers a similar range of relational functions, including comitative (26), benefactive (27) and ablative (28)8 in the Cantonese examples below:

(26) Kéuih tùhng yāt go pàhngyáuh góng s/he with one CL friend talk ‘He’s talking with a friend’.

(27) Ngóh tùhng léih ló yāt go I with you take one CL ‘I’ll get one for you’.

(28) Ngóh tùhng léih máaih yéh I with you buy things ‘I buy things from you’.

With regard to the structure [PP-V-NP] as in (23–28), it should be borne in mind that this is a typologically rare word order combination (Dryer 2003), leaving little

8. !ere are intriguing similarities between CCP and Tok Pisin long (see Smith 2002), which could result from a historical relationship and/or common grammaticalization paths.

!"#$%&''()


doubt as to the Sinitic origins of this feature in CCP. At the same time, preposi-tional phrases with long are also found following the verb, especially in English-language sources:

(29) I like werry much, do littee pidgeon long you. ‘I would very much like do to a bit of business with you’ (Downing 1838).

Another locative expression calqued on Cantonese is the construction with side, as in:

(30) bring that egg come thisee side ‘bring the eggs here’ (Instructor VI: 40)

(31) come Sydney side ‘(she) comes from Sydney’ (Instructor VI: 32)

Indeed the use of ‘side’ in this way is still commonly heard in Hong Kong English, as in Kowloon side ‘over in Kowloon’. !is is one of the few cases in which CCP ex-pressions have found their way into the present-day Hong Kong variety of English.

Serial verb constructions (SVCs) are frequent in CCP and follow a number of patterns (see Escure to appear). A well-known example is look see (calqued on Cantonese tái gin, ‘look-see’ = ‘look’):

(32) My wantchee look see counta ‘I want to check the accounts’ (Instructor VI.56)

(33) You look see dog no bitee you ‘Don’t let the dog bite you’ (Instructor VI.58).

Escure (to appear) also notes the frequency of directional serial verbs (34–36) based in particular on come and go. Directional serial verbs are also a feature of Sinitic languages, e.g.: daap fóchē làih, [take train come], ‘come by train’. SVCs can denote single as well as multiple events in CCP, e.g.:

(34) Bring come here ‘Bring it here’ (Instructor IV.43)

(35) What time you sendee tea come ‘When are you going to send the tea?’ (Instructor VI.16)

(36) Catchee one piece man go ‘Engage a man to go’ (Instructor IV.66)


4.5 Tense and aspect

!ere seems to be only one tense/aspect marker in CCP, the perfective marker hap (or hab), as in:

(37) my hap go court one time ‘I have been to Court once’ (Instructor IV.32)

(38) Coolie hap shutee alla window ‘Has the coolie shut all windows?’ (Instructor VI.53)

!e use of hab has a basis in Cantonese as well as English grammar. In interroga-tive sentences comparable to (38), Cantonese uses the existential verb yáuh ‘have’, as in (39):

(39) Yáuh móuh sāan saai chēung a? have not close all window SFP ‘Have you closed all the windows?’

Also seen in (38) is the quanti"er alla. In CCP alla can be used as a resumptive marker following a list of items (40–41), just like the equivalent Cantonese dōu (42):

(40) Green tea black tea alla hap got ‘I have both green and black tea’

(41) Two man alla same ‘we are both alike’

(42) Léuhng go yàhn dōu yāt yeuhng two CL person all one same ‘!ey are both the same’.

4.6 Wh-interrogatives in Chinese and Western sources

Interrogative sentences are especially revealing with regard to di&erences between Chinese and Western sources. English-language sources typically show wh-phras-es fronted as in English:

(43) What thing that Poo-Saat do? ‘What does Poo-Saat do?’ [Poo-Saat is a Chinese deity.] (Morrison 1807)

By contrast, the Instructor data show frequent use of wh-in-situ following the Chinese syntax:

!"#$%&''()


(44) You give what price ‘What price do you give?’ (Instructor VI.18)

!e Instructor shows variation between wh-in-situ (44) and wh-movement (43) which suggests that wh-movement is optional. In alternative pidgin renderings, the same wh-phrase appears variably in situ or fronted, even within the same dialogue:

(45) You wantchee how muchee? ‘How much do you want?’ (Instructor IV.54)

(46) How muchee more you wantchee? ‘What more do you want?’ (Instructor IV.54)

!e contrast between English-language sources (with consistent wh-fronting) and the Chinese phrasebook (with variation between wh-fronting and wh-in-situ) suggests that the CCP spoken by Chinese speakers di&ered systematically from CCP as used by native speakers of English and European languages.

4.7 Topic-comment and discourse structure

Another area where substrate in$uence can be discerned, as suggested by Shi (1991), involves topic-comment discourse structure. !e Instructor dialogues pro-vide just enough context for this feature to be observed. In the examples below, the topic, shown in square brackets, is the part that is being spoken of in the comment:

(47) [Good cargo] how can sellum cheap ‘How can good things sell cheap?’ (Instructor VI.11)

(48) [that pricee] he no sellum ‘He won’t sell at that price’ (Instructor IV.77)

As in Chinese, two kinds of topic can be distinguished: some are understood as arguments of the verb, like the object good cargo in (47); others are more loosely related to the predicate, like that pricee in (48) which is neither the subject nor object of sellum ‘sell’.

4.8 !e verb makee

A stereotypical feature of CCP is the use of makee (< make) before another verb:

(49) Go makee "ndee ‘Go "nd it’ (Instructor IV.45)

(50) makee catchee he


‘Seize hold of him’ (Instructor IV.77)

!e use of makee is a prominent feature of CCP which does not seem to re$ect substrate in$uence, since there is no apparent model for it in Cantonese. In these cases, makee merely indicates an action verb. !e phrase go makee $ndee (49) therefore represents a serial verb construction, in which directional go combines with the verb makee $ndee. In some cases, makee serves to create a verb from a noun, as in (51) where it combines with the Portuguese word conta ‘account’:

(51) my wantchee makee conta ‘I have to count them’ (Instructor VI.11)

(52) you hap long nother houso makee contract tea ‘Do you ever contract tea for anyone?’ (Instructor VI.19)

!is feature appears to be consistent across the various sources for CCP.

5. Conclusion

Based on the historical discussion and the grammatical data presented above, we can address some of the interpretations that have been o&ered about the formation of CCP. !e seeds of CCP were most likely sown in the early days of the Canton Trade in the 18th century along the Pearl River Delta, in particular in two settings: the harbor of Whampoa (Huangpu), which functioned as a sort of customs for the foreign merchandise, and the ‘factories’ of Canton (Guangzhou), i.e. the places of residence of the commercial representative of Western trading companies (Bolton 2003: 156; Martino 2003: 24; Van Dyke 2005).

Considering the structural parallels shown between Cantonese and CCP, it is safe to say that the CCP of "e Instructor exhibits substantial Cantonese in$uence. !e grammar of CCP can thus be accounted for "rst and foremost in terms of syn-tactic and semantic feature transfer from Cantonese. Since Cantonese is the domi-nant in$uence in the CCP of "e Instructor data, this could also explain why CCP is mostly isolating in morphology, without necessarily involving simpli"cation as a cognitive strategy. For example, zero-copula patterns may not be the result of faulty reproduction of English patterns, but rather an expression of Sinitic feature transfer. It is likely that further in-depth analysis of "e Instructor will reveal deep-er aspects of CCP grammar, together with more evidence of Sinitic substrate in$u-ence. !e features above should nonetheless su#ce to support the claims that the dominant features of CCP are Sinitic, including (i) noun classi"ers in the context [NUM/DEM-CL-NP]; (ii) zero copula; (iii) existential verbs; (iv) property verbs;

!"#$%&''()


(v) preverbal adverbs and PPs; (vi) serial verb constructions; (vii) topic-comment structure (Ansaldo 2009).

We have shown that in some features, the Chinese data contrast systematically with evidence from English-language sources, as in the case of wh-interrogatives where wh-in-situ appears only in the Chinese sources. Such data raise the ques-tion of variability between CCP as spoken by English-speaking and Cantonese-speaking users. Although the current analysis can not give a conclusive answer, the existence of a continuum of forms in$uenced by English or Cantonese gram-mar is considered likely, and the co-existence of more or less distinct lects whose poles were represented by a British and a Chinese register cannot be ruled out. If we think it unlikely that any stable pidgin was developed in the early days of the Canton Trade (for this view see Ansaldo 2009; Martino 2003), then we would tend to favor the latter interpretation. Be that as it may, the existence of the writ-ten Chinese language sources and the spread of CCP among Chinese clearly re-sulted in signi"cant in$uence of Sinitic grammar, vocabulary and pronunciation. Variation is likely to have been present all along, as suggested in Bolton (2003). Considering the limited number of speakers and the relatively short life-span of the ecologies in which it was spoken (Tryon et al. 1996), it seems unlikely that any great stability was ever achieved. We can thus envisage the evolution of CCP as follows:

In the earlier period, the position of inferiority the British found themselves in led them to simplify their language in order to try to communicate with the Chinese. Whether they did so intuitively or with the help of some perceived ‘knowledge’ of Chinese is unclear, but since the latter would have been di#cult to obtain, the former is more likely. In this phase, agency for CCP must also be attrib-uted to other Western traders since pidgin-evolution typically (but not invariably) involves third parties who use the language but are not speakers of the varieties that contribute to it. !is phase is related to the period of the Canton Trade.

As the British rose to power, their jargon increasingly became an attractive target for the Chinese. !e Chinese therefore set out to learn it and probably ex-panded it but it remained a clear trading jargon as is typical of ecologies where there is a strong desire to maintain social distance. !is phase most likely started in the factories and continued in other comparable environments.

It is thus possible to suggest that CCP was not the product of an encounter between two ‘high’ cultures: this has been supported by sociohistorical observa-tions about the inferiority of the British until the mid-19th century, and could explain why there is a body of CCP that looks more in$uenced by English than the data of "e Instructor. !e reason why we have stressed the segregation imposed, and partly achieved, by Chinese authorities, is not meant to suggest that contacts between Chinese and Western merchants were not signi"cant. But it is quite clear


that in an early phase, distance was kept between the parties, and that the Chinese were much less interested in the Western redhaired barbarians than the barbar-ians were in China. !erefore, the idea of CCP as initially pidginized by English speakers (and other foreigners) is a distinct possibility, as envisaged in Hall (1944; see also Ansaldo 2009). Van Dyke (2005: 80) reports at least one attestation of CCP being used by Western merchants in Canton as early as 1715, which accord-ing to him suggests that it was through merchants’ needs, rather than through the activities of linguists, that pidgin English developed. It was most likely developed by sailors and merchants who were not capable nor allowed to acquire Chinese mostly for informal trade (formal exchanges were most o%en taken care of by in-terpreters, but see Benson 2005 for an alternative view that direct contact between trading partners may have been more common than formerly thought).

However, CCP really grew in use compared with Portuguese only from the mid-18th century onwards, a time that we believe is more realistic for CCP to really come into being, as also argued in Martino (2003). Not only were Macau — and Makista — still dominant until that point, but commerce in the "rst half of the 18th century was still very restricted. Considering the super"cial and short-lived contacts between British and Chinese, further limited by the social distance imposed by the latter, CCP in the early years must have been very limited. A%er the Opium Wars, however, the situation changed dramatically: there was a great increase in commerce, in particular tea and silk; the contacts between Westerners and Chinese were no longer restricted to the Canton factories and Whampoa, but took place in several city-ports of the coast; and foreign merchants became ap-pointed as custom o#cers, while foreign companies appointed their own Chinese compradors (Fairbank & Goldman 1998: 203). In other words, the networks be-came more diverse, and as they increased in number and complexity, they favored the spread of CCP across China and, eventually, as far as Australia and the West Coast of the United States (Mühlhäusler & Baker 1996).

References

Adelaar, Alexander, & D.J. Prentice 1996. Malay: Its history, role and spread. In Stephen A. Wurm, Peter Mühlhäuser & Darrel Tryon (eds.), Atlas of languages of intercultural com-munication in the Paci$c, Asia and the Americas, 673–693. Berlin/New York: Mouton de Gruyter.

Anonymous. c. 1835. 紅毛通用番話 (hùhng mòuh tùng yuhng fàan wá — "e common language of the Red-haired Foreigners). Guangzhou.

Ansaldo, Umberto. 2007. Review of McWhorter, John, 2005, De"ning Creole. Oxford University Press. Journal of Pidgin and Creole Languages 22(1). 170–176.

!"#$%&''()


Ansaldo, Umberto 2009. Contact languages: Ecology and evolution in Asia. Cambridge: Cambridge University Press.

Ansaldo, Umberto & Matthews, Stephen. 2004. !e origins of Macanese reduplication. In Genevieve Escure & Armin Schwegler (eds.), Creoles, contact and language change: Linguistic and social implications, 1–19. Amsterdam/Philadelphia: John Benjamins.

Ansaldo, Umberto, Matthews, Stephen & Smith, Geo&. Forthcoming. !e Cantonese substrate in China Coast Pidgin. In Claire Lefebvre (ed.), Substrate in5uences in Creole languages.

Baker, Philip. 1987. Historical developments in Chinese Pidgin English and the nature of the relationship between the various pidgin Englishes of the Paci"c region. Journal of Pidgin and Creole Languages 2:2, 163–207.

Baker, Philip & Mühlhäusler, Peter. 1990. From business to pidgin. Journal of Asian Paci$c Communication 1(1). 87–115.

Benson, Philip. 2005. !e origins of Chinese Pidgin English: Evidence from Colin Campell’s diary. Hong Kong Journal of Applied Linguistics 10(1). 59–77.

Bisang, Walter. 1985. Das chinesische Pidgin Pidgin-English. Zürich: Universität Zürich.Bolton, Kingsley. 2003. Chinese Englishes: A sociolinguistic history. Cambridge: Cambridge

University Press.Boxer, Charles R. 1973. "e Portuguese seaborne empire. Hammondsworth: Pelican.Boxer, Charles R. (ed.). 1953. South China in the sixteenth century. London: Hakluyt Society.Chen, Ta. 1939. Emigrant communities in South China: A study of overseas migration and its

in5uence on standards of living and social change. Shanghai: Kelly & Walsh Ltd.Coates, Austin. 1966. Macao and the British: Prelude to Hong Kong 1637~1842. Oxford: Oxford

University Press.Downing, C. Toogood. 1838 [1972]. "e Fan-Qui in China in 1836–7. Shannon: Irish University

Press.Dryer, Matthew S. 2003. Word order in Sino-Tibetan languages from a typological and geo-

graphical perspective. In Graham !urgood & Randy LaPolla (eds.), Sino-Tibetan languag-es, 43–56. Richmond: Curzon Press.

Erbaugh, Mary & Yang Bei. 2006. Two general classi"ers in the Shanghai Wu dialect: A compari-son with Mandarin and Cantonese. Cahiers de Linguistique Asia Orientale 35(2). 169–207.

Escure, Genevieve. To appear. Is serialization simple? Evidence from Chinese Pidgin English. In !omas Klein & Nicolas Faraclas (eds.), Simpli$cation in pidgins and creoles. Amsterdam/Philadelphia: John Benjamins.

Fairbank, John K. 1953. Trade and diplomacy on the China Coast: "e opening of the treaty ports, 1842–1854. Stanford: Stanford University Press.

Fairbank, John K. & Goldman, Merle. 1998. China: A new history. Enlarged edition. Cambridge, MA: Harvard University Press.

Faraclas, Nicolas, Corum, Micah, Arrindell, Rhoda & Pierre, Jean O. 2007. Sociétés de cohabita-tion and the similarities between the English-lexi"er Creoles of the Atlantic and the Paci"c: !e case for di&usion from the Afro-Atlantic to the Paci"c. Paper presented at the 24th Meeting of the Society for Pidgin and Creole Linguistics.

Gunn, Geo&rey C. 1996. Encountering Macau: A Portuguese city-state on the periphery of China, 1557–1999. Boulder, CO: Westview.

Hall, Robert A.J. 1944. Chinese Pidgin English grammar and texts. Journal of the American Oriental Society 64. 95–113.

Hunter, William C. 1885. Bits of old China. London: Kegan, Paul, Trench and Co.Leland, Charles. 1892. Pidgin English sing-song. London: Kegan Paul.


Li, Michelle. 2011a. Chinese pidgin English and the origins of pidgin grammar. Doctoral dis-sertation. !e University of Hong Kong.

Li, Michelle 2001b. Origins of a Preposition: Chinese Pidgin English long and its Implications for Pidgin Grammar. Journal of Language Contact IV: 269–294.

Li, Michelle, Matthews, Stephen & Smith, Geo& P. 2005. Pidgin English texts from the Chinese-English Instructor. Hong Kong Journal of Applied Linguistics 10(1). 79–168.

Martino, Elena. 2003. Chinese pidgin English: Genesi ed evoluzione [Chinese Pidgin English: Genesis and evolution]. Tesi di Laurea. La Sapienza, Roma.

Matthews, Stephen & Yip, Virginia. 1994. Cantonese: A comprehensive grammar. London: Routledge.

Morrison, Robert. 1807–8. Unpublished journal in Council for World Mission archives, South China — Journals, Box 1.

Morse, Hosea B. 1926. "e chronicles of the East India Company trading to China, 1635–1834. 5 vols. Oxford: Clarendon Press.

Mühlhäusler, Peter, & Baker, Philip. 1996. English-derived contact languages in the Paci"c in the 20th century. In Stephen A. Wurm, Peter Mühlhäusler & Darrel T. Tryon (eds.), Atlas of languages of intercultural communication in the Paci$c, Asia, and the Americas, 497–522. Berlin: Mouton de Gruyter.

Mundy, Peter 1637. Itinerarium Mundii [!e travels of Peter Mundy]. Edited by R.C. Temple as "e Travels of Peter Mundy, in Europe and Asia, 1608–1667, volume 3 (1919). Cambridge: Hakluyt Society.

Nicol, John. 1822. "e life and adventures of John Nicol. Edinburgh: William Blackwood.Noble, Charles F. 1762. A voyage to the East Indies in 1747 and 1748. London: Becket and

Dehondt.Paviot, Jacques. 2005. Trade between Portugal and the Southern Netherlands in the 16th cen-

tury. In Ernst van Veen & Leonard Blussé (eds.), Rivalry and con5ict: European traders and Asian trading networks in the 16th and 17th centuries, 24–34. Leiden: CNWS publications.

Reinecke, John E. 1937. Marginal languages: A sociological survey of the creole languages and trade jargons. Yale University, Ann Arbor UMI.

Selby, Anne & Selby, Stephen. 1995. China Coast Pidgin English. Journal of the Hong Kong Branch of the Royal Asiatic Society 35. 113–141.

Shi, Ding Xu 1986. Chinese Pidgin English: Its origin and linguistic features. M.A. University of Pittsburgh.

Shi, Ding Xu. 1991. Chinese Pidgin English: Its origin and linguistic features. Journal of Chinese Linguistics 19(1). 1–40.

Shi, Ding Xu. 1993. Learning Chinese Pidgin English through Chinese characters. In Francis Byrne & John Holm (eds.), Atlantic meets Paci$c: A global view of pidginization and cre-olization, 459–464. Amsterdam: John Benjamins.

Siegel, Je&. 1990. Pidgin English in Nauru. Journal of Pidgin and Creole Languages 5(2). 157–186.Smith, Geo& P. 2002. Growing up with Tok Pisin: Contact, creolization, and change in Papua New

Guinea’s national language. London: Battlebridge.Smith, Geo& P. 2008. Chinese Pidgin English pronouns revisited. Hong Kong Journal of Applied

Linguistics 11(1). 63–76.Soothill, William E. 1925. China and the West: A sketch of their intercourse. Oxford: Oxford

University Press.Stoller, Paul. 1979. Social interaction and the development of stabilised pidgins. In Ian F.

Hancock (ed.), Readings in creole studies, 69–72. Ghent: Story-Scientia.

!"#$%&''()


Tamburello, Adolfo. 1983. La cultura occidentale nel Giappone Tokugawa, la mediazione olan-dese e russa nel 1800 [Western culture in Tokugawa Japan, Dutch and Russian mediation in 1800]. Il Giappone 21. 5–23.

Tong, King-Sing. 1862. "e Chinese and English Instructor, 6 vols. Guangzhou.Tryon, Darrel, Mühlhäusler, Peter & Baker, Philip. 1996. English-derived contact languag-

es in the Paci"c in the 19th century (excluding Australia). In Stephen A. Wurm, Peter Mühlhäusler & Darrel Tryon (eds.), Atlas of languages of intercultural communication in the Paci$c, Asia and the Americas, 471–495. New York: Mouton de Gruyter.

Van Dyke, Paul A. 2005. "e Canton trade: Life and enterprise on the China coast, 1700–1845. Hong Kong: Hong Kong University Press.

Whinnom, Keith. 1971. Linguistic hybridization and the ‘special case’ of pidgins and creoles. In Dell Hymes (ed.), Pidginization and creolization of languages, 91–115. Cambridge: Cambridge University Press.

Williams, Samuel W. 1836. Jargon spoken at Canton. Chinese Repository 4. 428–435.

+e African slave population of Portuguese IndiaDemographics and impact on Indo-Portuguese*

Hugo C. CardosoUniversidade de Macau

!is article is primarily concerned with quantifying the African(-born) popula-tion in the early Portuguese settlements in India and de"ning its linguistic pro"le, as a means to understand the extent and limitations of its impact on the emerging Indo-Portuguese creoles. Apart from long-established commercial links (includ-ing the slave trade) between East Africa and India, which could have facilitated linguistic interchange between the two regions, Smith (1984) and Clements (2000) also consider that the long African sojourn of all those travelling the Cape Route may have transported an African-developed pidgin to Asia. In this article, I concentrate on population displacement brought about by the slave trade. Published sources and data uncovered during archival research permit a characterisation of the African population in terms of (a) their numbers (rela-tive to the overall population), (b) their origin, and (c) their position within the colonial social scale. !e scenario that emerges for most territories of Portuguese India is that of a signi"cant slave population distributed over the colonial households in small numbers, in what is best described as a ‘homestead society’ (Chaudenson 1992, 2001). It is also made evident that there was a steady in$ux of slave imports well into the 19th century, and that the Bantu-speaking regions of modern-day Mozambique were the primary sources of slaves for the trade with Portuguese India.

1. Introduction

Paying heed to several authors’ warnings (e.g. Baker 1982; Arends 1995, 2008; Singler 1995), I assume that no attempt to explain the formation of a contact language is complete without accurate research into the territories’ social and

* I am very grateful to Umberto Ansaldo and two anonymous reviewers for their generous com-ments and suggestions. Any remaining shortcomings of this article are entirely my responsibility.

!"#$%&''()

92 Hugo C. Cardoso

demographic history. With respect to the Iberian-lexi"ed creoles, Jacques Arends voiced a common concern that ‘very little historical-linguistic work has been done, an unfortunate situation in view of the historical primacy of, especially, the Portuguese-lexicon creoles’ (Arends 2002: 50). In order to address this shortcom-ing, the present paper intends to contribute towards reconstructing the popula-tion make-up of the Portuguese-controlled territories in India at the beginning of colonisation (early 16th century) and therea%er, by assessing the size and distribu-tion of both the African-born population and their o&spring. !e data analysed in this study concerns primarily the northern territories of Portuguese India, in particular Daman and Diu, as well as Goa (see Map 1). Goa, Daman, and Diu were the territories under Portuguese rule for the longest period, from the early 16th century until 1961. Other regions were Portuguese strongholds for shorter periods (e.g. Cochin, Cannanore, Nagappattinam, Mylapor, etc.) and, due to the present scarcity of early documents, their demographic pro"le is considerably more dif-"cult to reconstruct.

Map 1. Main Portuguese settlements in Western India; 16th century.

!e African slave population of Portuguese India 93

!e presence of signi"cant numbers of Africans in the South Asian Portuguese settlements is well attested, not only in historical sources but also in the lyrics of Indo-Portuguese songs, which contain numerous references to Africans (see Schuchardt 1883, Dalgado 1921, Moniz 1925). !e cultural import of the African population is also clearly visible in a number of performing arts associated with communities from the former Portuguese-controlled areas of South Asia, par-ticularly music and dance styles such as the Goan mandó, the Sri Lankan and Mangalorean baila, and Sri Lankan cafrinha or ka!rinha (Jackson 1990, Jayasuriya 1996), known in the Moluccas as kafrínu and in Timor as kafrinia (Dalgado 1919). As far as language is concerned, however, it is a matter of some controversy how much the Indo-Portuguese and other Asian creoles owe to Africa. In addressing this issue, it is important to distinguish two main logical strands: a) the possibility that Indo-Portuguese was in$uenced by Portuguese-lexi"ed pidgin(s) or creole(s) transported from Africa; b) the possibility that, through the trade of African slaves into India, certain African languages entered the pool of early linguistic contribu-tors to the Indo-Portuguese creoles. !e "rst hypothesis interacts with the topic of this article only to the extent that African-developed pidgins and/or creoles may have been transported to India by displaced Africans, but I will survey some related proposals before turning to the discussion of the second hypothesis.

1.1 !e relationship between the Asian and African creoles

Ferraz (1987) addressed the possibility that the Portuguese-lexi"ed creoles of West Africa played a role in the development of their Asian counterparts, as he set out to assess the monogenetic hypothesis by comparing salient features in the Portuguese creoles of both continents. !e author identi"ed a set of similarities among the Asian creoles (e.g. the typical structure of the possessive construction, noun-mod-i"er word order, the future marker lo/logo and certain lexical items such as ada/ade[m]/adi ‘duck’), which are absent from those of Africa, therefore concluding that ‘there is indeed a degree of interrelatedness between the Portuguese Eastern Creoles, but that they are unrelated to those of West Africa’ (Ferraz 1987: 337). To justify the relative unity of the Asian creoles, Ferraz appealed to Dalgado’s (1917) notion of ‘partial reciprocal transfusion’: this hypothesis maintains that the Asian creoles were able to develop and maintain signi"cant similarities because the eco-nomical and political links between the territories facilitated strong cultural and population interchange.

While Ferraz’s typological dichotomy remains valid, Clements (2000) argues that he overstated the case by denying the possibility of any transfusion from Africa, since a number of similarities — though perhaps marginal — do suggest a degree of continuity between the African and Asian creoles. !e most striking

!"#$%&''()

94 Hugo C. Cardoso

example, for its apparent arbitrariness, is the fact that, in many creoles of both Africa and Asia, the base form of the verb ‘to go’ (typically vai/vay) is derived from an in$ected form in Portuguese while most other verbs draw on etymologi-cal non-"nite forms. Holm (1989) also draws attention to certain functional cor-respondences between grammatical items in Asian and African creoles, such as the completive use of the marker kaba. Clements (2000: 185–186) builds on this evidence to propose that:

para além dum pidgin Português geral, que se formou em África […] se formou, a partir do século XVI, outro pidgin português na Ásia. Este pidgin tinha caracterís-ticas em comum com o pidgin português geral mas, além destas, adoptou outros traços dos crioulos que se formaram na Ásia na primeira metade do século XVI.

[beside a general Portuguese pidgin formed in Africa […] another Portuguese pid-gin was formed in Asia a%er the 16th century. !is pidgin had certain character-istics in common with the general Portuguese pidgin but, besides those, adopted other traits from the creoles that formed in Asia in the "rst half of the 16th century.]

!e crews of Portuguese ships sailing the Cape Route (sailors, settlers, soldiers, mer-chants) were the most likely vehicle for the transmission of African-developed pid-gins/creoles. Smith (1984) voices a similar opinion, proposing that the Portuguese made use of structures encountered in African pidgins/creoles for communication in Asia. For the purpose of this paper, however, it is important to assess whether India-bound navigation was accompanied by the displacement of Africans who could have transmitted a Portuguese-lexi"ed pidgin or creole (see Section 4).

1.2 !e relationship between the Asian creoles and African languages

Given the in$ux of African slaves to the Portuguese territories of India, it is also worth considering whether any African languages other than the Portuguese-lexi"ed creoles may have contributed to the initial pool of linguistic features avail-able to the creators of Indo-Portuguese. !eir in$uence, if any, would have been particularly relevant at the historical moment of the creoles’ formation, which Clements (2000: 195) locates in the "rst half of the 16th century.

!e linguistic evidence for such a scenario is not particularly solid. !e most readily observable traces of African linguistic in$uence in Portuguese Asia are lexical. In his study of lexemes which were current among the ‘Portuguese’ (not necessarily Creole-speakers) in Asia, Dalgado (1919) identi"ed various words he traced back to African etyma,1 including:

1. Words derived from Arabic etyma were excluded from this study because, due to the wide geographical extension and in$uence of the language in the 16th century, it is o%en di#cult to


a. batuque ‘a type of drum’ (referentially equivalent to the Indic-derived term gumata): probably not derived from the Portuguese verb bater ‘to hit’, accord-ing to Dalgado (1919), but from similar terms used in East and West Africa to refer to a particular dance and the drum used to accompany it;

b. cacimba ‘dew, fog’: from Kimbundu kixima (ki’sima in other sources) ‘well’;c. calumba ‘a medicinal plant [Jateorhiza palmata]’: this plant is particularly as-

sociated with Southeast Africa, and the term derives from its name in local languages; e.g. Zulu –kalumuzi (Doke & Vilakazi 1964: 369);

d. machila, machira ‘palanquin’: the object is known by cognate terms in vari-ous East African languages; e.g. Makua machila, machira ‘palanquin’ (Prata 1990: 109);

e. mandó ‘a type of folk song and dance’: nowadays used mostly to refer to tradi-tional Konkani or Indo-Portuguese songs, Dalgado (1919) related the term to the word mandoa, found in Mozambican languages to refer to a local dance;

f. matomba, matomo ‘a tree [Parinarium excelsum]’: commonly believed to have been imported into the region of Goa from East Africa, where Dalgado (1919) recorded its name to be mutumbira;

g. pangaio ‘a type of boat’: cognate terms are widespread among both Indian and East African languages, but Dalgado (1919) placed its origin in Africa; e.g. Makua pangayo, pangaya ‘a type of boat’ (Prata 1990: 416);

h. pingo ‘grain of gold, gold bead’: unrelated to Portuguese pingo ‘drop’. Dalgado (1919) reported that the term píngu had a similar semantic value in (unidenti-"ed) languages of the Tete region of Mozambique.

Tomás (1992) added another word to the list which, signi"cantly, occurs in an Indo-Portuguese song from Daman recorded by Moniz (1925: 570, 571), a song speci"cally attributed to the ‘black’ section of the Catholic population:

i. muzungo ‘white man’: particularly common in East Africa, where cognates include Swahili mzunzu ‘white man, European’, Tsonga mulungu ‘white man, European’,2 or Makua musuku ‘white man, European, Portuguese’ (Prata 1990: 109).

ascertain whether they entered the Portuguese lexicon in Africa or elsewhere. As far as Eastern Bantu languages are concerned, in some cases many of them make use of cognate lexemes. !e examples provided here — from representative Southeast African languages (Makua, Swahili, Tsonga, and Zulu) — are simply meant to show their currency in the region, and should not be interpreted as de"nite etymologies for the words recorded by Dalgado (1919).

2. Cf. the 2003 edition of Dicionário Houaiss da língua portuguesa, entry muzúngu.

!"#$%&''()

96 Hugo C. Cardoso

Such lexical items could have been transported from the shores of Africa through various channels, including the trade contacts between dwellers of Portuguese-controlled India and dwellers of Portuguese-controlled Africa, population move-ments, as loans in Portuguese or through the languages of the displaced African slaves.

Scholars have long hypothesised that the creole spoken by Africans in Portuguese India may have di&ered from that of the remainder of the creolophone population. Spurred by an Indo-Portuguese song recorded in Goa, Dalgado (1921) proposed that a particular cafreal creole (from Ptg. cafre ‘black person’, cf. Note 3) was at some point spoken in the region by African slaves, a restructured variety which would have di&ered from the putative creole of the local Indo-Portuguese (see Section 4). !e hypothesis is unfortunately not well supported in Dalgado’s writings, nor is it clear whether his proposed cafreal creole was exclusive to Goa. One must be cautious not to over-interpret the fact that several Indo-Portuguese songs (viz. from Diu and Daman) are ascribed by their collectors exclusively to the African population; in fact, apart from some lexical peculiarities (see the dis-cussion of muzungo above), such songs display no structural traits which can-not be observed in the available non-cafreal data or modern-day varieties of Indo-Portuguese.

As a whole, the Indo-Portuguese creoles appear to have preserved no gram-matical traits clearly or solely attributable to (East) African languages (see Tomás 1992). Conversely, the typological link between these creoles and their immediate South Asian adstrates is well attested (see e.g. Smith 1984, Clements 1996, Cardoso 2009). !e remainder of this article will explore documentary evidence concerning the sociodemographic characteristics and linguistic pro"le of the (early) African slave population in Portuguese India, which may shed light on the extent and limi-tations of their ethnolinguistic contribution to the Indo-Portuguese communities.

!e delineation of precise "gures must out of necessity rely on archival sources, but there are several obstacles for this type of research. To begin with, a great deal of the earliest sources and records kept by the Portuguese has been lost, notably in the destruction of Lisbon by the 1755 earthquake but also on account of their antiquity. Moreover, not enough archival work has been done concerning the spe-ci"c issue of the African slave trade to India. Despite these shortcomings, a num-ber of references have so far been unearthed that provide direct demographic data concerning African slaves. However, other sources also clarify that the presence of Africans in India predates the arrival of the Portuguese. It is therefore crucial to distinguish between African migration to India brought about by Portuguese colonial activities in the region, and migration prior and/or unrelated to it.


2. Africans in India prior to European colonial presence

At present, the term Siddhi is used overarchingly to refer to all sections of the Indian population who, though considered native to India by virtue of their long history in the subcontinent, ultimately have their origins in Africa. Nonetheless, it would be a mistake to trace the presence of all Siddhis back to just one historical period or channel of migration. !e institution of slavery in South Asia predates the arrival of Vasco da Gama (see Chauhan 1995: 228), as does the trade in African slaves. !e transport of African slaves to the Indian subcontinent is probably as old as any other sort of trade between the two regions, with Pankhurst (2003) providing evidence that slaves were being carried from the territory of modern-day Ethiopia to Western India as early as the 1st century AD. !e documentary evidence is much more robust a%er the 13th century. At the time, slaves were im-ported from Abyssinia via Arabia, most eventually converting to Islam and ending up particularly in Gujarat and the Gulf of Cambay or further south on the Malabar coast and Ceylon. In the period preceding the European colonial expansion, this type of trade across the Indian Ocean seems to have been "rmly dominated by Arab traders. No estimate of the numbers involved in this trade is available, but the presence of Africans in India and their roles are very well attested. In ear-ly documents, a number of di&erent terms are used to refer to African people, some of which are quite revealing as to their (perceived) provenance. !e term Habshi has traditionally been related to Habash, the Arabic name of Abyssinia, modern Ethiopia (cf. Jayasuriya & Pankhurst 2003b: 8). !e term Siddhi or Siddi, also attested in old documents, is less revealing than Habshi in that it carries no geographical connotation; it has instead been traced back to Arabic Saiyid ‘lord’. Finally, another common term found in archival texts is Ka6r,3 the Arabic word for ‘in"del’.

!e import of African slaves through these channels did not halt a%er the Portuguese, French, and British made their presence felt in India. According to Pankhurst (2003: 200–201), the slave trade from Ethiopia even brie$y intensi-"ed a%er 1527 as a result of military expeditions taking place in Ethiopia itself. With respect to its longevity, precious information can be retrieved from an 1853 document in which the consul of Britain in Muscat referred to active slave trade links between the Sultanate of Zanzibar and the Indian states of Kutch and Katiawar.4

3. Among the Portuguese, the term was cafre. It is interesting that, according to Lodhi (1992), the Siddhis of Diu are still known as Kafaras.

4. Both Kutch (or Kachchh) and Katiawar (or Kathiawad) integrate modern-day Gujarat. !e Katiawar region is presently known as Saurashtra and refers to the area immediately surrounding

!"#$%&''()

98 Hugo C. Cardoso

Even though many Africans were taken to India in a condition of servitude, it is important not to underestimate their historical role as seafarers and traders in the Indian Ocean. Ethiopians were particularly active in this activity up to the 7th century. Baker (1996) observes that, even a%er the demise of their maritime power, Ethiopians were particularly prized as slaves, in particular in the Islamic societies surrounding the Indian Ocean, and o%en granted positions of great re-sponsibility. One needs to concede that, in these societies, ‘the status of slaves […] was clearly very di&erent from that of Africans subsequently enslaved by Christian nations in the Caribbean and elsewhere’ (Baker 1996: 645). Medieval India was no exception in this respect. As a matter of fact, historical records narrate strik-ing instances of the social ascension of Africans in Muslim and Hindu India. In his 16th-century tour of India, Ibn Batutta related the case of Badr, an Abyssinian who rose to become the governor of Alapur (North India). A lineage of 14th-century governors of Jaunpur (North India) is also said to have been initiated by a Habshi slave. 15th-century annals contain several examples of African interven-tion in Indian politics, such as a Habshi-led coup in Bengal (Northeast India), their hold on governor positions in the Deccan (West India), involvement in a Bijapuri (South India) strife for the throne, military leadership in Goa, etc. (see Chauhan 1995, Pankhurst 2003). !e most impressive example of Siddhi ascen-sion is perhaps the unique history of the island-fort of Janjira and nearby Danda-Rajpuri, on the Konkan coast south of Bombay and very close to the ancient Chaul — where modern Korlai Indo-Portuguese originated (see Clements 1996). From the late 15th or early 16th century onwards, the Siddhis e&ectively held control of the island and contiguous territory for approximately two centuries.5

In most of these cases, the notoriety of people of African ancestry was a product of their military performance. As early as 1333–1334, Ibn Batutta reported having come across a corps of "%y Abyssinian soldiers on a ship in Gandhar, in Gujarat (quoted in Pankhurst 2003: 192). In the early 16th century, Tomé Pires’ Suma Oriental described how the rulers of the Cambay region employed large num-bers of Abyssinian soldiers (Pires 1978: 136, 163). According to a contemporary

Diu. Not surprisingly, the Saurashtra region is home to an important Siddhi population, as is Diu itself (see Lodhi 1992).

5. Various 16th-century Portuguese sources contain references to mjilic dastur (Malik Dastur), ruler of Janjira who ‘he abixij esc̃pauo DellRey tam homrrado casy como cada huũ destes’ [is an Abyssinian slave of the King, as honored as any of these (other lords)] (Pires 1978: 209). While Pankhurst (2003) identi"es the 17th and 18th centuries as periods of decline in overall Siddhi in$uence in India, their presence in Janjira was still going strong despite the Maratha campaign of 1659 under Shivaji. !ey remained key players in the area, entering treaties with the British in Bombay, and also with the Portuguese against Shivaji in 1670 (Chauhan 1995: 58).


observer, the Gujarati scholar Haji ad-Dabir, the Moghul force which occupied Gujarat in 1572 included some 700 Habshi horsemen.

It is precisely in a military context that the "rst encounters between Siddhis/Habshis and the incoming Portuguese seem to have happened: the Portuguese had to "ght Habshi forces in order to occupy some of their territories, notably those belonging to the Sultan of Cambay. Daman is described as a stronghold of the Habshis, who continued to challenge Portuguese rule even a%er their conquest of the city in 1559, and their presence has also been reported in Diu when the city was occupied, in 1535 (Chauhan 1995). Crucially for our present topic, both Daman and Diu were to become loci of Indo-Portuguese creoles. !e presence of a previous African population in these regions, therefore, calls for constraint in tak-ing modern Siddhi demographics as indicators of the number of slaves imported during the Portuguese period. Considering that the Portuguese had the habit of baptizing their slaves, it is quite telling that the Gujarati Siddhis, including those of Daman and Diu, are presently Muslim and not Christian (but see Section 3.3.2, and Note 13). By contrast, 41.9% of the Siddhis of Karnataka, which is next to Goa, are Catholic (1980s data, in Lobo 1984: 40), while others are Muslim and Hindu (Lodhi 1992: 83). !e history of the Karnatakan Siddhis is probably some-what di&erent from that of the Gujarati Siddhis; Chauhan (1995: 231) suggests that forced conversion to Christianity in Goa drove many slaves to escape into the neighbouring territories, where they would have been integrated into older Siddhi communities. And yet, this type of escape is also documented for the northern territories, chie$y in Daman: a 1695 treaty between the Portuguese Viceroy and the neighbouring Chauthia Raja made provisions for the return of runaway slaves to the city’s authorities (Chauhan 1995: 233). !is issue brings us to the topic of the Portuguese importation of slaves, which is dealt with in the following section.

3. Africans in India as a result of European colonial activities

Africans were employed in Portuguese India in a number of functions, the most obvious of which was domestic service. Mandelslo, a traveller in Goa, reported in 1638–1639 (quoted in Chauhan 1995: 230):

Most of the Portuguese have many slaves of both sexes whom they employ not only about their persons but also upon any other business they are capable of, for what they get, comes into the master. Whence it comes that handsome wenches are sought a%er, to be employed in the selling of fruits and such commodities as the Portuguese send to market, to the end their beauty might draw in customers. !eir keeping as to diet stands them in very little. !e children born between

!"#$%&''()

100 Hugo C. Cardoso

slaves belong to the master, unless the father will redeem them within eight or ten days a%er they are born.

!ere were also African soldiers among the Portuguese military forces. Although the Estado da Índia’s6 armies included Portuguese, Indians, and mestiço soldiers as well, Pinto (1992: 82) noticed that ‘[t]he Estado’s military establishment depend-ed on African slaves as soldiers in all its territories’, more so ‘than did any other European colonising nation’.

Africans were also essential to crew sea-going vessels, to such an extent that Linschoten (1598) reported maritime work to be one of the main occupations of the ‘free’ Habshi:

!ese Abexiins and Arabians as are frée doe serve in al India for saylers and sea faring mẽ, with such marchants as saile from Goa to China, Iapon, Bengala, Mallaca, Ormus, and all the Oriental coast.

It is unclear whether the Abexiins mentioned by Linschoten were manumitted Africans brought by the Portuguese as slaves, given that this term was not the most current among the Portuguese to refer to their African slaves. !e following sub-section will explore in more detail the referential value of various terms employed in Portuguese sources.

3.1 Terminology in Portuguese documents

In old Portuguese texts, various terms are used to refer to Africans, but in some cases there is overlap with the words the same authors use to refer to the local Indians. !e terms preto ‘black’ and negro ‘black’ are particularly ambiguous. In o#cial documents such as censuses, the ambiguity was resolved by employing the terms nativo and natural ‘native’ for Indians, while cafre was reserved for Africans. Notice, for instance, the following entry from the Goan Inquisition records of 1651, in which a Mozambican-born (from the Rios de Cuama region, see 3.2.) was classi"ed as a cafre:

Jorge, em gentio Hiamata cafre natural dos rios de Cuama, e morador nesta cida-de por arrenegar da nossa santa fée, e blasfemar contra a pureza de nossa senhora, passandose a seita de Calvino e a de Mafamede em terra de mouros. (Arquivo Nacional da Torre do Tombo; Conselho Geral do Santo Ofício [Inquisição de Goa], nº 33 nº 1)

6. Estado da Índia was initially the generic designation for the whole of the Portuguese-controlled territories in Asia as well as East Africa (see Morais 1997: 9).


[Jorge, Hiamata in gentile, cafre born at the rivers of Cuama, and dwelling in this city [Goa] for denying our holy faith and saying blasphemy against the purity of our lady, adopting the sect of Calvin and that of Mohammed in a moorish land.]

To complicate matters further, one cannot immediately interpret the term escravo ‘slave’ as referring to an African — though in the vast majority of cases it did — because there were other sources of slaves for the Portuguese, including certain regions of East Asia and even India. While in Goa, Linschoten (1598) witnessed a speci"c type of bondage involving Indian individuals:

Also in time of povertie or dearth the fathers may sell their children, as it hap-pened in my time, that there was such a dearth, and scarcitie of victualls in the "rme lande, and countries bordering uppon Goa, that the men of India came to Goa (and other places where the Portingales are resident) to sell their children, in great numbers, and for small prices, to buy them victuals. […] and some came with their wives and children to o&er themselves to bee slaves, so that they might have meate and drinke to nourish their bodies. And because the Portingales have tra#que in all places, (as we have béene in many) it is the cause why so many are brought out of all countries to be solde, for the Portingales doe make a living by buying and selling of them, as they doe with other wares.

Jayasuriya & Pankhurst (2003: 11) further observe that the Portuguese obtained slaves from the Muscat Arab forces they vanquished o& Diu in 1670, and Pinto (1992) mentions that at times they would trade supplies for slaves when famine struck the Coromandel coast (southeast India). According to o#cial slave libera-tion registries, the Cartas de Alforria aos Escravos, most of the 753 slaves manu-mitted in Goa between 1682 and 1759 were Indian, and only 6 (all female) were African. !ere is also documentary evidence of slave trade being carried out from emporia in Java, Makassar on the island of Sulawesi, Japan and China (see Pinto 1992: 18). It is clear, however, that the volume of trade in all these areas never ap-proached that from Africa. !erefore, the next section will be dedicated to assess-ing the ethnogeographical origin of African slaves.

3.2 Origins of the slaves

!e slave trade to India involved various routes at di&erent periods of time, and these relied on di&erent areas of Africa for the supply of slaves. Whereas the most-ly Arab-controlled trade seems to have sought slaves in Northeast Africa, there is solid evidence that the Portuguese conducted their activities further south, in the Bantu-speaking area that nowadays roughly corresponds to Mozambique. Notice, for instance, the inscription on the map of the Island of Mozambique in the English-language edition of Linschoten’s travels (Linschoten 1598: 8–9), which

!"#$%&''()

102 Hugo C. Cardoso

reads ‘!e discription of the Islandes, and Castle of Mozambique, lyeinge up-pon the borders of Melinde, rich in Ebanwood, "ne Goulde, and Ambergrize, frõ whence many Slaues are caried into India’. Even more explicitly, Linschoten stated that ‘From Mosambique great numbers of these Ca&ares are caried into India’.

In the 17th century, observers such as Pyrard de Laval and Crooke were also very speci"c in their observations that, by then, most of the Africans in Goa were from Mozambique or Mombasa (Pankhurst 2003: 215). Later, in 1738, the Portuguese viceroy complained of the lack of manpower in the following terms: ‘por não terem vindo cafres de Mossambique, e haver falta delles na terra, faltará o serviço, que no trabalho do conves costuma fazer esta gente’ [since no blacks came from Mozambique, and there is a lack of them in the land, the service these people usually do on the deck will be lacking] (quoted in Boxer 1984).

In several instances, the Portuguese merchants would not capture the slaves themselves. Consider the following observations of Linschoten (1598) concerning the Ca!ares of Mozambique:

they doe commonly make warre one against the other, and some of them eate mens $esh, and some there are also that eate it not, but such as deale with the Portingals. When they take any man prisoner in the warres, they sell him to the Portingales, or exchange and barter him for Cotton linnen, and other Indian wares.

Trade with African slavers was carried out in speci"c emporia, with the implica-tion that even if a certain place of origin is documented, the bulk of the slaves may have originally come from further a"eld. Nonetheless, it is possible to pin-point with some accuracy the general area of the African coast in which slaving activities were carried out by the Portuguese. A decree issued by the Portuguese authorities in 1816, although relatively late, is rather explicit in that respect: it rules that the only slaver vessels allowed should be those ‘que se destinarem a fazer o Commercio de Escravos nos pórtos da Cósta Oriental de Africa, comprehendidos entre o 10.mo e 25.o graus de Latitude Austral’ [destined to trade slaves in the ports of the Eastern Coast of Africa between the 10th and 25th degrees of lati-tude south].7 !ese latitudes limited the Portuguese possessions on the East coast of Africa and correspond quite closely to the northern and southern borders of present-day Mozambique.

Several documents record the Rios de Cuama, which refers to the Zambezi river delta in Mozambique, as the native place of slaves (cf. Inquisition record transcribed above). !e 1855 registry of slaves in Daman (see 3.3.2. below) is particularly revealing in that all of those who had not been born in India were

7. Arquivo Nacional da Torre do Tombo; Junta do Comércio, Maço 62, Caixa 204.


registered as natives of the Rios de Cuama.8 An Indo-Portuguese song from Daman also provides some ethnogeographical information concerning the provenance of the African population. Notice that all references (Sofala, Inhambane, Sena, Macua) refer back to the Mozambican territory. !e relevant words are empha-sised in the following transcription, but not in the original:9

1º 3ºCabelos torcidos Seus beiços compridosCafarinho despido Seus olhos torcidosToda gente fala Rosto de rabanaTem cafre de Sofala Tem cafre de Inhambane

CORO 4ºHuê, huê, huê, balha com igual huê Beicinhos furadosHuê, huê, huê, cabelos torcidos huê Seus dentes limadosHuê, huê, huê, festa de Natal huê Cafarinho pangaio

Pinchando na praia2º

5ºCafarinho tem pireitoTorcido e bem feito Todos assim dizemBalhando na rua Chapado narizComo cafre de macua Cabeça pequena

Tem cafarinho de Sena

Medeiros (2003) gives ample evidence that the Portuguese slavers provided slaves to the French possessions on the Mascarenes. From the 18th century onwards, this trade proceeded as an act of smuggling, given that slave trading with the French was prohibited at that time so as to concentrate the slave ‘resources’ in the exploi-tation of the colonies of Brazil. An ethnographic study of the African population of the Mascarenes, carried out in the 1840s by Eugène de Froberville (reported in

8. !is region was believed to be rich in mineral resources, and as such saw a thrust of interest from the Portuguese Crown and settlement e&orts in the 1660s and 1670s (see Ames 1998). !e Senas nowadays form the majority in this coastal area, with Nyungwe being spoken upriver into the Tete province.

9. !e version transcribed here was published by an anonymous author (1982) in the Revista da Academia da Língua e Cultura Portugesa as provided by the Mendonça family of Daman. In Moniz (1925), we "nd the transcription of the same song with some minor di&erences. !e most relevant di&erence for this study is that the last verse of the song apud Moniz is Tem cafarinho de Somaliz instead of Sena.

!"#$%&''()

104 Hugo C. Cardoso

Medeiros 2003: 73), identi"ed various ethnicities, including the Makua and natives of Inhambane in Mozambique. Medeiros expands on the issue and observes that the slaves transported to the Mascarenes from Mozambique ‘came mainly from the ethnic groups, Marave (Cheua-Nianja), Senas, Maconde, Macua, Machona and Yemvane (Inhambanes)’ (Medeiros 2003: 71).10

Although one must also admit the presence of slaves from other parts of Africa, all documentary evidence is unanimous in locating the primary source of slaves transported to Portuguese India (and Portuguese Asia) within the Mozambican territory. !e following sub-section will analyse the available records in order to quantify the African slave population in the territory, with particular reference to Goa and the northern territories of Western India (see Map 1).

3.3 Demography

One of the challenges to de"ning the demography of African slaves in the Estado da Índia results from the scattered and long-standing nature of this political unit, which means the scarce documentary sources available refer to di&erent territo-ries and di&erent historical periods. Rather than being able to fully reconstruct the demographic make-up of speci"c settlements throughout their history, one is mostly le% with information on transregional trends. One of the most interesting conclusions we can draw from early 17th-century descriptions of the Portuguese settlements, notably those by Bocarro (1635), is that slaves were present even in the smallest among them. With reference to Agaçaim (in the vicinity of Bassein, modern-day Vasai), Bocarro wrote the following:

a um quarto de légua dela, está uma povoação de uma rua somente, lançada de norte a sul, que tem trinta moradores portugueses, […] com poucos escravos

[a quarter of a league away from it, there is a village with one street only, drawn from north to south, which has thirty Portuguese dwellers, […] with few slaves]

However, di&erent territories hosted di&erent numbers of slaves, with Goa taking the lead as the great slaving emporium of Portuguese India. I have therefore opted for a separate discussion of Goa and the northern territories (the Província do Norte).

10. Madagascar also provided slaves for the trade, but for the most part it remained peripheral to the Portuguese: Newitt (2003: 84) recognised that the practice of acquiring slaves in Madagascar for India was sporadic and of minor signi"cance. From the 17th century onwards, Madagascar itself began requiring slaves from the mainland, not only for domestic use but also for re-ex-port, and therefore turned into a slaving platform onto which converged slaves captured in Mozambique; this fact certainly ties in with the observation that certain communities known as Makoi have preserved a distinct identity in Madagascar and are reported to have spoken Makua until the 1920s (Medeiros 2003: 65).


3.3.1 Goa!e "gures concerning the slave population of Goa, reconstructed from docu-ments as diverse as travel accounts, business records, and censuses, are truly im-pressive. Consider the description of Goa in Linschoten’s travel rotary, based on his visit to the territory in the 1580s:

likewise they have running about them, many sorts of captives and slaves, both men and women, young and old, who are daylie sould there, as beasts are sold with us, where everie one may chuse which liketh him best, everie one at a cer-taine price. !ere are some married Portingales, that get their livings by their slaves, both men and women, whereof some have 12, some 20 and some 30, for it costeth them but little to kéepe them. !ese slaves for money doe labour for such as have néede of their helpe, some fetch fresh water, and sell it for money about the stréetes: the women slaves make all sorts of confectures and conserves of Indian fruites, much fyne néedle worke, both cut and wrought workes, and thẽ their master send the fairest and the youngest of them well drest up with their wares about the stréetes to sell the same, that by the neatnes & bewtie of the said women slaves, men might be moved to buy, which happeneth more for the a&ection they have to the slaves & to ful"ll their pleasure with them, then for any desire to the conserves or néedle workes: for these slaves never refuse them, but make their daylie living thereby, and with the gaines they by that meanes bring home, their masters may well main-taine them. !e Portingales and Mesticos in India never worke, if they doe, it is but very little, and that not o%en […]: but most of them have their slaves to worke in their shops, and the masters when they walke up and downe the stréetes, goe as proud-lie as the best.

While Linschoten claimed the number of slaves possessed by single owners range dbetween 12 and 30, Bocarro (1635) proposed a Goan average of 10. Pinto (1992: 26) quotes Gemelli-Careri’s report of his visit to Goa in 1695, in which he stated ‘[t]here are also an abundance of Cafres and Blacks; for there are Portuguese that keep thirty or forty, and the least six or twelve’. It would probably be a mistake to believe that all Portuguese settlers possessed as many slaves or any slaves at all (but notice Linschoten’s claims concerning the Portuguese cra%speople). It is not unlikely that such concentrations of slaves were the prerogative of the wealthiest. Equally high "gures feature in records pertaining to the religious orders in Goa. !e most striking of these concerns a complaint by the residents of the Convent of Santa Mónica in Old Goa (unknown date) that their 120 slaves were insu#cient; the document added that even single individuals in the community could have ‘"%een or twenty female slaves, or 26 women and girls, while a juiz ordinario or a

!"#$%&''()

106 Hugo C. Cardoso

desembargador held 85 female slaves… and some rich ladies over three hundred’ (Pinto 1992: 27).

In addition to being the wealthiest Portuguese city in India, Goa functioned as the main platform for the diaspora of slaves throughout the Estado da Índia and, as such, high import "gures should not come as a surprise.11 !e territory’s abundance of African slaves deeply impressed most of its visitors in the 16th century, who never failed to comment on the issue. One point to be remembered is that large concentra-tions of slaves in single households or institutions, such as those described above, appear to have been exclusive to Goa. !e evidence from most of the northern terri-tories, as we shall see in the following sub-section, suggests a starkly di&erent picture.

3.3.2 "e Província do NorteAt its peak, in the 16th/17th century, the Província do Norte ‘Northern Province’ comprised the stretch of land between Daman and Korlai (including Bassein and Bombay, see Map 1) and also Diu across the Gulf of Cambay. Bassein was the leading settlement of the entire area, a city of considerable extension and mag-ni"cence with the epithet Cidade dos Fidalgos ‘City of the Noblemen’. With regard to Bassein, Bocarro advanced a tentative average of slaves per Portuguese settler:

[o]s casados que haverá nesta cidade, brancos serão quatrocentos os mais deles "dalgos, com pretos cristãos virão fazer seiscentos e todos terão uns pelos outros três escravos cada um. (Bocarro 1635)

[the casados who dwell in this city must be four hundred whites, most of them noblemen, and with the christian blacks it should make six hundred, and all of them would have in average three slaves each]

Bocarro’s calculations would result in a slave population approaching 1800, in 1635. However, it is important to realise that the "gures for Bassein may not be automatically extended to the rest of the Província. Comparable data has fortu-nately been unearthed for other northern settlements, in particular for Daman. According to Moniz (1923), the number of slaves in Daman, in 1660, was esti-mated at around 600, although the overall population of the territory is not given. On the other hand, customs registries indicate that Daman was an entrepôt for the import and distribution of slaves at least until 1828 (despite the 1773 decree which had formally abolished slavery in Portugal). Pinto (1992: 29,30) quotes one

11. !e slaves purchased in Goa were not only meant for the Estado but for a variety of buy-ers. !e 1777 correspondence between a Frenchman and a Goan agent reveals that, within two months, two ships from Mauritius had purchased 160 and 140 slaves respectively, and also that in September that year a ship from Mozambique arrived in Goa transporting some 700 slaves (Pinto 1992: 38).


Table 1. Registry of slaves in Daman, 1855.Male Female Birthplace Age

1 X Africa 132 X Daman 83 X Daman 64 X Daman 45 X Africa 356 X Daman 17 X Daman 148 X Africa 689 X Africa 3510 X Africa 4811 X Daman 2012 X Daman 1413 X Daman 1614 X Daman 815 X Africa 7516 X Africa 6017 X Daman 1318 X Africa 1419 X Africa 3020 X Daman 1421 X Daman 522 X Daman 323 X Daman 5 months24 X Africa 6025 X Africa 3526 X Daman 327 X Africa 2828 X Africa 6029 X Daman 2030 X Daman 1831 X Daman 3832 X Africa 2533 X Daman 1534 X Daman 8

!"#$%&''()

108 Hugo C. Cardoso

such registry from the year 1828, according to which ‘48 male and female African slaves landed in Daman port in a week from 27th September to the 4th of October’. !e author further quotes numbers retrievable for the period between 1837–1839, which turn out much lower than this (10–15 in 1837; 8–10 in 1838; 5–7 in 1839). Nonetheless, an 1855 census of slaves12 still reported 43 slaves in Daman. !e data contained in the registry is given in Table 1.

!e owner of each slave is also given in the registry, revealing that they were distributed over just 12 households in groups never larger than 6; half of the households had only 1 or 2 slaves. !e distribution of slaves per household is pre-sented in Table 2.

!e census further indicates that all African-born slaves were originally from the Rios de Cuama region (see Section 3.2.). It is also interesting to notice that 46.5% of the slaves were locally-born, all of them under the age of 38; these were probably the children of the registered African-born slaves. One implication, not without consequence, is that children born to slave parents were considered slaves themselves; this ties in with previous accounts which make it clear that this was a long-standing practice in the Portuguese-controlled territories: recall Mandelslo’s 17th-century account quoted above, and consider !evenot’s late 18th-century description of Daman, which stated that ‘the Portuguese have slaves there of both

12. Goa Historical Archives; Registo dos Escravos da cidade de Damão, doc. 2979.

Table 2. Distribution of slaves per household, 1855.Master nr. Nr. of slaves

1 12 63 54 25 36 17 58 19 2

10 111 312 4

TOTAL 12 34


sexes, which work and procreate only for their Masters, to whom the Children belong, to be disposed of at pleasure’ (quoted in Pinto 1992: 28).

Another of the territories constituting the Província do Norte was Diu, for which detailed but scattered demographic information can be found. !e earliest available reference to the slave population is found in Bocarro’s (1634) description, which reads:

os casados portuguezes, que vivem oje nesta cidade de fora da fortaleza, são cin-coenta e nove, avendo já sido muitos mais: são pobres, pelas ditas causas, mas ain-da assim tem huns por outros cincoenta e nove escravos que possão tomar armas, as quaes tem alguns de cabides de lanças muy bastantes pera brigarem.

[the Portuguese casados living in this city outside the fortress today are "%y-nine, although they used to be more numerous: they are poor, for the previously men-tioned reasons, but among them still have some "%y-nine slaves capable of taking up arms, some having enough spears to wage war.]

!e 1792 census of the Christian population counted 104 slaves without an age break-up, all but 7 of whom in the two parishes within the city walls (Sé Matrix and S. !omé). Table 3 provides the census data.

Table 3. Census of the Diuese Christians, 1792 (adapted from Pinto 1992: 31).Parishes Prior of

cathedralVicars Bene--

ciariesMen> 14

Men< 14

Women> 12

Women< 12

Slave men

Slavewomen

Sé Matrix 1 2 64 27 114 16 27 43 294S. !omé 1 103 23 67 17 15 12 238St. AndréExtramuros

1 15 12 25 6 3 4 66

TOTAL 1 2 2 182 62 206 39 45 59 598

According to this census, the 104 slaves made up 17.4% of the Catholic population. !is means at least part of the slaves in late 18th-century Diu were baptised and possibly integrated to some extent into the Catholic community.13 A later census, in 1842,14 surveyed the entire population of the territory, which totalled 9373 peo-ple (see Cardoso 2009:67). While this "gure was distributed according to religious groups (cristãos ‘Christians’, gentios ‘gentiles [i.e. Hindus]’ and mouros ‘Moors [i.e. Muslims]’), there was a 274-strong ‘racial’ category — negros (também chamados escravos) ‘blacks (also called slaves)’ — representing nearly 3% of the whole. It

13. Certain members of the Catholic community in modern-day Diu still have some typically African physical features. Upon inquiry, this fact was easily recognised and con"rmed by their peers.

14. Goa Historical Archives; População de Dio, doc. 2997.

!"#$%&''()

110 Hugo C. Cardoso

is striking that the 1842 ‘slave’ population was more than twice that of "%y years earlier. !e discrepancy, however, is probably due to criterial variation. In 1792, only the Christians had been surveyed; in 1842, it is not clear whether the resident (Muslim) Siddhis were included in the negros category or the mouros section, but the former seems most likely. Otherwise, one would have to assume a slave popula-tion much larger than anything glimpsed from previous records and also that only a minority of the slaves employed in Diu were Catholic — which is unlikely given the habit of the Portuguese to baptise their slaves (cf. Pinto 1992). Mass import of slaves in this period is also highly unlikely, given that, by then, Diu was already in deep commercial decline (Cardoso 2009). It is therefore warranted to postulate that the slave population of Diu in 1842 would have made up 1–2% of the population and comprised of mostly Christians. In 1855, a slave registry sheet, similar to the one produced for Daman (see Table 1), was also prepared for Diu15 but never "lled in. Nonetheless, it constitutes the latest known mention of slavery in the territory.

4. Discussion

!e data presented above suggests the various territories of Portuguese India dif-fered substantially with respect to the demographics of African slaves and their so-cial embedding. 16th-century "gures lead us to believe that Goa was an outlier in that it contained a larger than average slave population; two important facts are a) that the mobility of these slaves was rather high (with Goa serving as an import-export platform for the slave trade), but also b) that the average of slaves per colonial household in Goa was still much higher than those advanced by the same authors for e.g. Bassein. Goa never developed a plantation culture of the type that was to appear later in the Caribbean and elsewhere. Instead, most of the slaves were employed in domestic tasks, in variable concentrations ranging from small to rather high num-bers but always in relative proximity to their masters. Slaves were not con"ned to the colonial houses, if one is to believe Linschoten’s remarks that they were sent out to service others in the community and sell their masters’ wares in the market.

!is type of social organisation corresponds to what Chaudenson (1992: 93) calls the société d’habitation ‘homestead society’, a phase (in certain cases prepara-tory of a société de plantation ‘plantation society’) characterised by some proximity between slaves and masters which, nevertheless, did not compromise the latter’s social domination. On the other hand, some of Goa’s institutions, such as convents and seminaries, were able to congregate vast numbers of African slaves. !ese cases are peculiar in that they conform less with the prototypical homestead phase

15. Goa Historical Archives; Registo dos Escravos de Diu, doc. 2981.


described by Chaudenson, while not allowing for the type of social detachment between slaves and colonists typical of the Atlantic-type plantation settings. In this light, Dalgado’s (1921) hypothesis of a cafreal creole, although not unequivocal, is interesting; he proposed to trace a presumably Goan song with Indo-Portuguese characteristics back to the Convent of Santa Mónica in Old Goa, partly on the intuition that the high concentration of slaves made it a likely locus for the use of a creole language among Africans. See Section 3.3.1. for a concurrent description of this convent as housing an unusual number of African slaves.

!e situation concerning the Província do Norte is more straightforward than that of Goa. While 16th-century documents identify the presence of slaves in most Portuguese settlements, they consistently mention lower numbers (though less so in Bassein). Here too, the picture that emerges is that of a homestead setting, in which a few households congregated a small number of slaves (see Table 2 above). Crucially, in Chaudenson’s (1992) view, the homestead phase, when combined with a di&erential in linguistic prestige, favours the abandonment of the language(s) of the slaves in favour of the language(s) of the masters. In the case of the Província, one might postulate that this assimilation was towards Indo-Portuguese; one fac-tor in support of this scenario is that the African slaves were mostly converted to Catholicism (before or upon arrival in India) and consequently entered a well-de-"ned community whose bulk was not European but local and Eurasian: the speak-ers of Indo-Portuguese. While no data earlier than the 18th-century is available to support this claim, the Diuese Catholic demographics of 1792 given in Table 3 — in which slaves make up 17.4% of the total — and the general census of 1842 — in which the proportion of non-European to European Catholics is of more than 13 to 1 (see Cardoso 2009:66–67) — are very revealing.

One of the questions raised in the introduction, in connection with Clements’ (2000) suggestion of some continuity between the African creoles and their Asian counterparts, is whether African slaves could have been vehicles of (restructured) Portuguese into India, i.e. whether they spoke Portuguese (standard, pidgin, or creole) upon arrival in India. To introduce this issue, consider the 1778 letter from a French settler in Mahé to an agent in Goa in which he requested ‘to buy a ka6r of the very best type. Age about nineteen or twenty, knowing a little Portuguese, intelligent and having already if possible, done some cooking for which work he is intended’ (quoted in Pinto 1992: 38).

!is document can give a hint as to the knowledge of Portuguese on the part of African slaves. !e request reveals that "nding a slave in Goa with some knowledge of Portuguese was not impossible, but it also suggests this could not be taken for granted. It may not have been too common to come across a newly imported Mozambican slave who already possessed any knowledge of Portuguese. On the one hand, it has been shown that most of these slaves were war captives,

!"#$%&''()

112 Hugo C. Cardoso

from regions where Portuguese settlement was minimal. On the other hand, con-cerning Mozambique, Gonçalves (2004: 230) points out that ‘[a]lthough the pres-ence of Portuguese people in the country dates back to the end of the 15th-century (1498), the di&usion of the language was practically nil during the "rst four centu-ries of colonisation’ by virtue of the colony’s peripheral position in face of India or Brazil. !e Rios de Cuama region itself was not as deeply a&ected by the late 17th-century settling thrusts as envisaged (see also Ames 1998). Linschoten (1598) also hinted at some lack of interest on the part of the Portuguese authorities when he discussed the spread of Christianity among locals in Mozambique:

!ere are some of them that are become Christians since the Portingales came thether, but there is no great paines taken about it in those coũtries, because there is no pro"te to be had, as also that it is an infectious and unwholesome countrie.

Given this scenario, any eventual knowledge of Portuguese (pidgin) could only have been picked up during the journey to India, in which case one must see the ships’ crew, not the slaves, as the main link to restructured Portuguese as spoken in Africa — if at all. !is begs the question of the fate of the slaves’ native Bantu languages a%er their settlement in India. Two sociohistorical factors seem to have contradictory implications for the linguistic assimilation of the slaves: on the one hand, it is clear that the in$ux of Mozambican slaves was constant until at least the early 19th-century; on the other hand, it was also shown that a signi"cant number of slaves were actually born in India of slave parents and possibly raised in an Indo-Portuguese (Catholic) context. In the particular case of Portuguese India, linguistic assimilation appears to have taken the upper hand, though not without the retention of certain Bantu-derived terms which appear to have been a trademark of the population of African origin.16 !erefore, even though it makes the social picture more complex, the recognition of an important population dis-placed from Africa does not compromise the classi"cation of the Indo-Portuguese creoles as essentially endogenous.

References

Ames, Glenn J. 1998. An African Eldorado? !e Portuguese quest for wealth and power in Mozambique and the Rios de Cuama, c. 1661–1683. "e International Journal of African Historical Studies 31 (1). 91–110.

16. !ere are scattered (uncon"rmed) suggestions that some Siddhis retain, to this day, either some form of ritual language with African undertones or African in$uences in their speech. Lodhi (1992: 83), for instance, claims that ‘in Daman and Diu they speak Gujarati with many Swahili/Bantu words and phrases’.


Anonymous. 1982. Cancioneiro. Revista da Academia da Língua e Cultura Portugesa, II (3). 29–32.

Arends, Jacques. 1995. Demographic factors in the formation of Sranan. In Jacques Arends (ed.), "e early stages of creolization. Amsterdam/Philadelphia: John Benjamins.

Arends, Jacques. 2002. !e historical study of creoles and the future of creole studies. In Glenn Gilbert (ed.), Pidgin and Creole linguistics in the twenty-$rst century. New York: Peter Lang.

Arends, Jacques. 2008. A demographic perspective on creole formation. In Silvia Kouwenberg & John Victor Singler (eds.), Handbook of Pidgin and Creole Studies. Oxford: Basil Blackwell.

Baker, Philip. 1982. "e contribution of non-Francophone immigrants to the lexicon of Mauritian Creole. London: University of London doctoral dissertation.

Baker, Philip. 1996. !e potential for the development of Arabic-based and other contact lan-guages along the maritime trade routes between the Middle East and China, from the start of the Christian era. In Stephen A. Wurm, Peter Mühlhäusler & Darrell T. Tryon (eds.), Atlas of languages of intercultural communication in the Paci$c, Asia, and the Americas, vol. II (1), 637–672. Berlin/New York: Mouton de Gruyter.

Bocarro, António. 1635. Livro das plantas de todas as fortelezas, cidades e povoações do Estado da India Oriental.

Boxer, Charles Ralph. 1984. !e Carreira da Índia (ships, men, cargoes, voyages). In Charles Ralph Boxer (ed.), From Lisbon to Goa, 1500–1750: studies in Portuguese maritime enter-prise. London: Variorum Reprints.

Cardoso, Hugo. 2009. "e Indo-Portuguese language of Diu. Amsterdam: Universiteit van Amsterdam doctoral dissertation [pub. Utrecht: LOT].

Chaudenson, Robert. 1992. Des îles, des hommes, des langues. Paris: L’Harmattan.Chaudenson, Robert. 2001. Creolization of language and culture. London: Routledge.Chauhan, R.R.S. 1995. Africans in India: from slavery to royalty. New Delhi: Asian Publication

Services.Clements, Joseph Clancy. 1996. "e genesis of a language. Amsterdam/Philadelphia: John

Benjamins.Clements, Joseph Clancy. 2000. Evidência para a existência de um pidgin português asiático. In

Ernesto d’Andrade, Dulce Pereira and Maria Antónia Mota (eds.), Crioulos de base lexical portuguesa. Braga: Associação Portuguesa de Linguística.

Dalgado, Sebastião Rodolfo. 1917. Dialecto Indo-Português de Negapatão. Revista Lusitana 20. 40–53.

Dalgado, Sebastião Rodolfo. 1919. Glossário Luso-Asiático. Coimbra: Imprensa da Universidade.Dalgado, Sebastião Rodolfo. 1921. Berço duma cantiga em Indo-Português. Revista Lusitana

22. Separata.Doke, C.M. & Bilakazi, B.W. 1964. Zulu-English dictionary. 2nd ed. revised. Johannesburg:

Witwatersrand University Press.Ferraz, Luís Ivens. 1987. Portuguese creoles of West Africa and Asia. In Glenn Gilbert (ed.),

Pidgin and Creole languages. Honolulu: University of Hawaii Press.Gonçalves, Perpétua. 2004. Towards a uni"ed vision of classes of language acquisition and

change: arguments from the genesis of Mozambican African Portuguese. Journal of Pidgin and Creole Languages 19 (2). 225–259.

Holm, John. 1989. Pidgins and creoles (2 vols). Cambridge: Cambridge University Press.Jackson, Kenneth David. 1990. Sing without shame. Oral traditions in Indo-Portuguese Creole

verse. Amsterdam/Philadelphia: John Benjamins.

!"#$%&''()

114 Hugo C. Cardoso

Jayasuriya, Shihan de Silva. 1996. Indo-Portuguese songs of Sri Lanka: !e Nevill Manuscript. Bulletin of the School of Oriental and African Studies 59 (2). 253–267.

Jayasuriya, Shihan de Silva & Pankhurst, Richard (eds.). 2003a. "e African diaspora in the Indian Ocean. New Jersey: Africa World Press.

Jayasuriya, Shihan de Silva & Pankhurst, Richard. 2003b. On the African diaspora in the Indian Ocean. In Shihan de Silva Jayasuriya & Richard Pankhurst (2003a).

Linschoten, Jan Huygen van. 1598. "e Voyage of John Huyghen van Linschoten. London: Hakluyt Society.

Lobo, Cyprian. 1984. Siddis in Karnataka. Bangalore: Jesuits of Bangalore.Lodhi, Abdulaziz Y. 1992. African Settlements in India. Nordic Journal of African Studies 1 (1).

83–86.Medeiros, Eduardo. 2003. Contribution of the Mozambican diaspora in the development of

cultural identities in the Indian Ocean Islands. In Shihan de Silva Jayasuriya & Richard Pankhurst (2003a).

Moniz, António Francisco. 1923. Notícias e documentos para a história de Damão: Antiga provín-cia do norte. Bastorá: Tipogra"a Rangel.

Moniz, António Francisco. 1925. !e Negroes and St. Benedict’s feast. In "e mission $eld. "e diocese of Damaun, 570–572. Bombay: published by S. R. Santos.

Morais, Carlos Alexandre de. 1997. Cronologia geral da Índia Portuguesa. 1498–1962. Lisbon: Editorial Estampa.

Newitt, Malyn. 2003. Madagascar and the African Diaspora. In Shihan de Silva Jayasuriya & Richard Pankhurst (2003a).

Pankhurst, Richard. 2003. !e Ethiopian diaspora to India: !e role of Habshis and Sidis from medieval times to the end of the eighteenth century. In "e African Diaspora in the Indian Ocean. In Shihan de Silva Jayasuriya & Richard Pankhurst (2003a).

Pinto, Jeanette. 1992. Slavery in Portuguese India (1510–1842). Bombay: Himalaya Publishing House.

Pires, Tomé. 1978 ("rst published 1550). Suma Oriental. Coimbra: Imprensa da Universidade de Coimbra.

Prata, António Pires. 1990. Dicionário Macua-Português. Lisboa: Instituto de Investigação Cientí"ca e Tropical.

Schuchardt, Hugo. 1883. Kreolische Studien III. Über das Indoportugiesische von Diu. Sitzungsberichte der Kaiserlichen Akademie der Wissenscha#en zu Wien (philosophisch-historische Klasse), 103. 3–18.

Singler, John Victor. 1995. !e demographics of creole genesis in the Caribbean: A comparison of Martinique and Haiti. In Jacques Arends (ed.), "e early stages of creolization, 203–232. Amsterdam/Philadelphia: John Benjamins.

Smith, Ian. 1984. !e development of morphosyntax in Sri Lanka Portuguese. In Mark Sebba & Loreto Todd (eds.), York papers in Linguistics. York: University of York.

Tomás, Maria Isabel. 1992. A presença africana nos crioulos portugueses do Oriente: o crioulo de Damão. In Ernesto d’Andrade & Alain Kihm (eds.), Actas do Colóquio sobre Crioulos de Base Lexical Portuguesa, 97–108. Lisbon: Colibri.

Vestiges of etymological gender in Malacca Creole Portuguese*

Alan N. BaxterUniversidade de Macau

!is paper is concerned with the nature of etymological gender in$ection in Malacca Creole Portuguese and its place in the diachrony of the language. !e discussion considers pairs of items with etymological gender in$ection and hu-man referents. Speaker intuitions of the word-pair vitality are assessed by word recognition and judgments of the acceptability of test words in speci"c syntactic contexts. !e items surveyed are initially contemplated as a type of derivational morphology. Subsequently, the question is posed as to whether this etymological gender morphology could have a contextual dimension, and it is observed that some limited traces thereof may be found. !e incorporation and retention of etymological gender in$ection is attributed to second language input to the cre-ole, involving early second language lexical and phrasal development, in$uenced by a combination of communication pragmatics, salience, and potential Indian substrate features. It is suggested that such input could be both an early dia-chronic occurrence or could have arisen through an ongoing diglossia involving a wider lectal grid that may have existed until the 19th century.

1. Introduction

Within the Asian varieties of Creole Portuguese,1 the Southeast Asian and East Asian varieties of Malacca, Macau, and Tugu (Batavia) are those that approximate the most to McWhorter’s (2005) controversial view of the creole language proto-type. !ey are highly analytical in structure and do not display the in$ectional

* Research for this paper was supported by Universidade de Macau grant RG-UL/07–08S/Y1/BA01/FSH. I would like to thank three anonymous reviewers, Umberto Ansaldo, Hugo Cardoso, Armin Schwegler, and Mauro Fernández for their helpful comments on earlier dra%s. All remaining shortcomings are my own.

1. For economy, the term Creole Portuguese (CP) will be used as a synonym of Portuguese lexically-based Creole throughout this paper.

!"#$%&''()

116 Alan N. Baxter

verbal morphology that has long been associated with many Indo-Portuguese creoles, such as Daman CP (Dalgado 1906, Clements & Koontz-Garboden 2002) or Korlai CP (Clements 1996). Nevertheless, as with all the Portuguese lexically-based creoles, the East and South East Asian varieties have etymological vestiges of Portuguese bound morphology.2

!is paper focuses on the nature of gender morphology in Malacca Creole Portuguese and its implications for the restructuring of Portuguese NP morpholo-gy in the development of this language. !e paper incorporates comparisons with certain Iberian-based creoles of the Asian region, and some beyond the region. However, a comparative study of bound morphology in Iberian-based creoles in general is a task far beyond the scope of the present work.

2. +e language

Malacca Creole Portuguese (MCP), now endangered, has some 750 speakers in Malacca, West Malaysia, and is the last vital variety of Creole Portuguese in the Southeast and East Asia region (Baxter 2005). It belongs to the group of Southeast Asian Creole Portuguese varieties referred to as Malayo-Portuguese (Holm 1989) and its roots lie in the Portuguese conquest of Malacca in 1511. During the 16th and 17th centuries until its conquest by the Dutch in 1642, Malacca was a key port in a trade and administration network of Portuguese establishments extending from Goa to the Moluccas and beyond to Macau.3 !is network saw a constant tra#c of Portuguese and their various camp followers, and traders of other origins.4

!e setting for the development of MCP was highly complex. It would have involved the importation of both foreigner talk and pidgin models from India and

2. !e term bound morphology is used to refer collectively to morphemes that are bound to a root, as opposed to morphemes that occur independently.

3. Portuguese establishments in Southeast and East Asia in the sixteenth and seventeenth cen-turies took three main forms: a mere trading post (feitoria), a forti"ed stronghold (fortaleza) and, in the case of Malacca and Macau, port-cities (Villiers 1987: 38).

4. !e European Portuguese were always a small minority in their Asian outposts and they drew heavily on local peoples, free and enslaved, in maritime trade and military endeavours. Missionary activities created local Christian populations of Portuguese cultural and linguistic orientation. Furthermore, there were very few Portuguese women who ventured out to Asia in the early centuries, and large-scale miscegenation was the rule (Baxter 1988, 1996; Clements & Koontz-Garboden 2002: 197). Also signi"cant, in terms of social cohesion and control, was the creation of a casado class, of European Portuguese o#cially married to local women, which helped develop Luso-Asian populations loyal to Portugal.

Vestiges of etymological gender in Malacca Creole Portuguese 117

beyond, a local genesis, and the eventual introduction of creole models from India (Baxter 1988, Holm 1989).5 Subsequently, there may have been further in$uence on the creole through continuity of trade with Asian ports where Portuguese and Creole Portuguese communities existed, a point that will be discussed in Section 9. !e importance of the Indo-Portuguese connection is supported by typological features shared across many of the Asian Portuguese-based Creoles. Such items include widespread accusative-dative marking, Tense-Mood particles or auxilia-ries derived from Portuguese logo ‘presently’ and há de ‘will’ (Ferraz 1988: 350), and a post-nominal genitive (Hancock 1973, Ferraz 1988, Baxter 1988).6 However, the grammar of MCP is strongly in$uenced by its primary and secondary local substrates, Malay and Hokkien (Baxter 1988, 1996, Hancock 1975). Because of its strong similarity to its two o&shoots, Macau CP7 and Tugu CP8 (Baxter 1996,

5. !e conquest of Malacca by the Portuguese in 1511 occurred a very short while a%er their arrival in India. Since Portuguese contacts in the Upper Guinea region began in the second half of the 15th century, it is likely that pidgin models from Africa would have reached Asia. It might be assumed that Portuguese arriving in India a%er lengthy voyages that included sojourns in Africa would have had some familiarity with pidginized Portuguese and the use of foreigner talk (Clements & Koontz-Garboden, 2002: 199). Nevertheless, Clements (1992: 88) sensibly points out that it is a reasonable assumption that Pidgin Portuguese in the Atlantic region had several di&erent sources and di&erent birthplaces at di&erent times and it is our view that the same may also be said for the Asian region. While input from the Atlantic region is very likely, local contexts would also have led to pidginization. !e existence of contact varieties of Portuguese formed in 20th century Brazil, such as that of the Xingu indigenous reservation in Mato Grosso (Emmerich 1984, 1993), demonstrates that contact situations and foreigner talk can produce similar results without any link to a prior pidgin.

6. !omaz (1985) draws attention to the signi"cant role of Indian merchants in Portuguese Malacca, and suggests that, at the beginning of the seventeenth century, most of the local Christian converts were of Indian origin.

7. Malacca was the forward base for the foundation of Macau in 1555/1557. Macau was closely linked with Malacca until the Dutch conquest of the latter in1642, and maintained signi"cant trade links with it subsequently. Macau CP appears to have had considerable input from MCP, with which it bears strong similarities on all levels. However, it has always been in contact with Portuguese and has been decreolizing since the nineteenth century, while also su&ering from lan-guage shi% in the Macanese creole community, initially towards Portuguese and, more recently, Cantonese. Macau CP is now highly endangered. With few speakers in Macau, it is better repre-sented in the diaspora, especially in the Macanese communities in San Francisco and Vancouver. For an overview of the development and structure of Macau CP, see Holm (1989) and Baxter (2009).

8. Tugu CP has its roots in the introduction to Batavia of slaves and free persons originating from Portuguese territories and trading posts conquered by the Dutch during the mid and sec-ond half of the 17th century. !ese included Ceylon, enclaves on the Malabar and Coromandel

!"#$%&''()

118 Alan N. Baxter

Holm 1989: 294–298, Schuchardt 1891), it seems reasonable to assume that MCP to a good extent resembled its present analytical form before 1642 (Hancock 1973, 1975, Baxter 1988, Holm 1989).

3. +e basis for the study

!e presence of some bound morphology in MCP is not a novelty. It is "rst men-tioned in Rêgo (1942 [1998]: 63), in relation to vestigial gender in nouns and adjec-tives. Subsequently Hancock (1969) brie$y mentions de-verbal adjectives formed via the a#x -du and vestigial gender in adjectives. Vestigial gender morphology in nouns and adjectives is also referred to by Baxter (1988: 49–50, 59) and Tomás (2004: 140–141, 145). None of these previous works explore the topic of bound morphology in any detail. However, recent work on the lexicon of MCP (Baxter & de Silva 2004) has revealed that bound etymological morphemes, although a minor feature of the language, are somewhat more signi"cant than previously be-lieved and, curiously, gender morphemes appear to have a degree of vitality.

Research over the last two decades has now rendered uncontroversial the notion that creole languages can have bound morphology (e.g. Plag 2001, 2005, DeGra& 2001, Baptista 2002, Bakker 2003, Becker & Veenstra 2003, Holm 2005), challenging the traditional assumption and claims regarding the paucity of trans-mission of in$ectional and derivational morphology from lexi"er to pidgin or to creole. !e presence of derivational morphology seems relatively uncontroversial (Plag 2001), aside from the question of its dating: early input, input via long-term or recent coexistence with the lexi"er? On the other hand, the presence of in$ec-tional morphology in creoles raises important questions concerning the nature and signi"cance of the morphological categories involved (Kihm 2003).

However, at the same time, it is evident that the discussion of the status of sub-areas of bound morphology in creoles still requires in general a richer set of data from which to draw substantial conclusions. In this respect, research on the MCP lexicon prompted a re-evaluation of some of Bakker’s assertions regarding the lack of gender morphology in the nominal domain of Romance-lexi"ed creoles:9

In contrast to the lexi"er, there is no trace of grammatical gender in creole nomi-nal in$ection or adjectival in$ection […] !ere is only very rarely a gender di&er-ence in the personal pronouns, and that is only in starkly decreolized languages.

coasts and, signi"cantly, Malacca. Tugu CP as documented in Schuchardt (1891) does not di&er greatly from MCP.

9. Crowley (2008: 81–82) makes a similar comment.


[…] Adjectives in creole lexi"ers display gender […] None of this is found in cre-oles. Valdman (1978: 149–152), however, discusses a few gender-distinct nouns and adjectives in French creoles, adding that this is exceptional in creoles. In pid-gins, however, gender is sometimes marked on nouns. (Bakker 2003: 16–18)

Data from MCP may make an interesting contribution to this debate. Bearing in mind the current discussion of bound morphology in creole languages, the aims of the present paper are to:

a. evaluate the signi"cance and workings of gender morphology in MCP and its relation to language contact involving Portuguese, focusing on the potential role of L2 data in the development process;

b. assess whether such morphological data could permit diachronic dating.

!e paper is organized as follows: A%er a brief account of the descriptive frame-work and the data sources and procedures, we turn to a consideration of the status and vitality of gender morphology in MCP. Subsequently, we pose the question of whether MCP gender morphology could be considered as involving intra-phrasal or inter-phrasal relations, and consider the possible motives for the retention of this morphology. Finally, the paper brie$y addresses the topic of whether there was diglossia involving Creole Portuguese and Portuguese following the Dutch takeover of Malacca in 1642. Such a possibility could place vestigial gender mor-phology into the realms of bilingual or L2 acquisition of acrolectal features.

4. Framework for the discussion

!e foregoing discussion will assume the traditional prototypical distinction be-tween derivational and in$ectional morphemes. !us, in$ection is viewed as the obligatory form that a lexeme assumes for grammatical purposes, without any change to the lexical category or to the core meaning of the lexeme (Arono& & Fudeman 2005: 160, Booij 2007: 112). On the other hand, derivation is viewed as involving the optional creation of new lexemes typically with di&erent core lexical meanings, yet with or without a change of lexical category (Arono& & Fudeman 2005: 160). In comparison with derivation, in$ection is considered to be more productive, and is viewed as functional with any root possessing the ap-propriate syntactic and semantic characteristics (Arono& & Fudeman 2005: 161, Bybee 1985: 11). In$ection is also said to always produce a predictable, more gen-eral meaning, whereas derivation has more semantic content (Bybee 1985: 99). Not withstanding, in$ection and derivation are better viewed as part of a cline, as proposed by Bybee (1985) and others who have demonstrated amply that the

!"#$%&''()

120 Alan N. Baxter

traditional de"nitions are porous: some derivational items may have grammatical qualities, just as some in$ections may have derivational qualities.

Current discussions of pidgin and creole morphology (cf. Kihm 2003, Roberts & Bresnan 2008) have found the notions of inherent in5ection and contextual in-5ection proposed by Booij (1995, 2007) to be useful. Inherent in$ection is de"ned by Booij (1995: 2) as ‘the kind of in$ection that is not required by the syntactic con-text, although it may have syntactic relevance’. It is determined by speaker choice concerning the information to be conveyed, rather than by the syntactic context (Booij 2007: 104). In contrast, contextual in$ection pertains to the representation of relationships between words in syntactic structure (Booij 2007: 104–112). !us, for example, in the nominal domain, inherent in$ection concerns the representa-tion of number, whereas contextual in$ection concerns case and agreement. Here too we will assume this distinction, which we believe may be especially useful for contemplating cases wherein the distinctions between in$ection and derivation tend to blur.

Kihm (2003) has observed that there appears to be a preference in creole lan-guages for inherent in$ection over contextual in$ection. Recently, Plag (2008) has proposed an interesting account of the motives for this preferential pattern in terms of his restatement of the Interlanguage Hypothesis10 and Pienemann’s (2000, 2005) Processability !eory. From Plag’s perspective, an absence of contex-tual in$ection would be explained by the assumption that creoles are convention-alized early-stage interlanguages. On the other hand, structures of more advanced interlanguage would be the result of prolonged access to the lexi"er language. Plag (2008) points out that Processability !eory proposes stages in the develop-ment of processing procedures that correspond to di&erent sequential levels of morphological and structural development which may overlap. An initial stage of a complete lack of morphology is followed by a lexical stage in which inherent in$ection (part of a category procedure, in Processability !eory terminology) is operational. Following this stage, two types of contextual in$ection are imple-mented in sequence. Firstly, a procedure of intra-phrasal information exchange, such as NP-internal agreement. Subsequently, a phase of inter-phrasal informa-tion exchange begins, comprising an ordered sequence of sub-types and involving, for example, SV agreement. !us, following Plag’s (2008) reasoning, the nature of a#xal morphology in creoles might be viewed as re$ecting di&erent moments of morphological development present in early stage interlanguage. Aspects of the Interlanguage Hypothesis and Processability !eory will be considered when evaluating the MCP data.

10. !e notion that creoles are derived from an input of early interlanguage, originally was proposed by Andersen (1980).


5. Data and procedures

!e data to be discussed are adjectives and nouns containing etymological in$ec-tional morphemes, which were observed initially in personal narratives and tradi-tional stories recorded at various moments in the period 1980–1985. !ese words were incorporated into a word-list to which additions were subsequently made on the basis of similar recordings made at various times in the 1990s and again in the 2000s. Prior to their inclusion in Baxter & de Silva (2004), the vitality of all the items discussed in this paper was assessed by a cross-section of native speakers of MCP. Subsequently, a pilot study of the vitality of etymological derivational and in$ectional morphemes was conducted in 2006.

In April 2009 an additional survey was conducted directly addressing speaker intuitions in relation to words containing etymological gender morphemes. Fi%een informants were tested, selected from households of the Malacca Portuguese Settlement,11 the neighborhood where the majority of MCP speakers reside. !e households have been known by the researcher since 1980 to be traditional settings for MCP use. !e informants are distributed in two broad age-groups: (i) 30–55 (4 female, 4 male); (ii) 65–75 (3 female, 4 male). !e prime aim of the study was not variationist, and was not directly concerned with the maintenance or loss of the morphology, but rather with determining whether the morphology has functional vitality among speakers from traditional creole-speaking households. Hence, special attention was not given to the constitution of discrete age cells, or to controlling fur-ther extralinguistic factors such as, for example, education, employment, or social network facts. !ese are considerations for future work on lexical retention in MCP.

Two types of testing procedures were used, a word-recognition test and con-textual tests, each requiring approximately 45 minutes to apply. !e tests were ap-plied on separate occasions, with individual informants. !e word-recognition test evaluated a list of 26 noun pairs and 13 adjective pairs with etymological male and female gender morphemes in MCP, as attested in previously collected corpora. To this list were added 80 common words, comprising nouns of three types ([+hum], [−hum, +anim] both devoid of etymological gender yet also including male and female referents, and [−anim]), adjectives devoid of etymological gender, adverbs and verbs. !e list was scrambled so as to separate pairs of test items. !e words were read to the informants, who were asked to signal if they recognized the word

11. !e Portuguese Settlement, now part of urban Malacca, was created in 1933 as a resettle-ment village for poorer creoles from Malacca town and from the seaside community at both Praya Lane and Banda Kanu to the immediate south of the town, (see Baxter 1988, 2005 for details).

!"#$%&''()

122 Alan N. Baxter

and to provide the equivalent in English or Malay.12 !e recognition scores are shown in Table 1 in Section 6, below.

!e contextual tests involved the evaluation of the vitality of etymological gen-der items within the context of particular structures. Four lists were prepared of pairs of sentences involving, respectively, male and female potential ‘controllers’ yet including only the potential ‘target’ item with male etymological in$ection.13 !ese sentences were of four types:

i. !ose containing an NP with a ‘controller’ head N and a ‘target’ adjective with etymological in$ection:

a. ńgua máchu prigasózu one male lazy.m ‘a lazy male/fellow’

b. ńgua fémi ?prigasózu one female lazy.m ‘a lazy female/woman’

ii. !ose containing an NP with a ‘controller’ head N and a ‘target’ adjective with etymological in$ection, but with marked word-order in a focusing struc-ture that permits dislocation of the adjective to initial position, prior to the determiner:

a. prigasózu ńgua máchu lazy.m 1 male ‘a lazy male/fellow’

b. ? prigasózu ńgua fémi lazy.m 1 female ‘a lazy female/woman’

12. Prior to beginning, the informant was given a couple of examples and it was carefully ex-plained what was required. !e words were supplied quickly to the informant, in order not to allow re$ection. Generally, two (sometimes three) responses could be obtained within one minute.

13. !e terms controller and target are borrowed from Corbett (1991, 2006). In agreement sys-tems, the controller is the word that determines the morphological agreement on another word, the target. However, in the present context, where no such agreement is assumed, these terms are used in order to evaluate whether the choice of a certain ‘target’ form (adjective or noun) with etymological gender in$ection is or is not determined by a ‘controller’, whether semanti-cally or morpho-syntactically.


iii. Attributive predications, containing a ‘controller’ subject NP and a predicate with a ‘target’ attributive adjective with etymological in$ection:

a. aké máchu (bomóng) prigasózu dem male good-good lazy.m ‘that male/fellow is very lazy’

b. aké fémi (bomóng) ?prigasózu dem female good-good lazy.m ‘that female/woman is very lazy’

iv. Attributive predications, containing a ‘controller’ subject NP and a predicate with a ‘target’ attributive N with etymological in$ection:

a. yo sa kanyóng kuzinyéru 1s poss eldest brother cook.m ‘My eldest brother is a cook’.

b. yo sa súsi ?kuzinyéru 1s poss eldest sister cook.m ‘My eldest sister is a cook’.

In type (iv) sentences, only ten female referent nouns were tested. !ese were words registered during the preparation of Baxter & de Silva (2004) and again ver-i"ed in the pilot study in 2006 as being consistently distinguished by informants as requiring female referents (see the nouns in bold in Table 1, Section 6 of this paper). !ese previous assessments had been done only by testing the recognition of words in isolation.

To each of the four lists of sentences were added an equal number of sentences of each structural type, half with male and half with female referent nouns in ‘con-troller’ positions, but in this case, the ‘target’ position featured only adjectives and nouns that do not contain etymological in$ections, as for example bésta ‘stupid (m or f)’, chádu ‘clever (m or f)’, méstri di skóla ‘school teacher (m or f)’. Subsequently, all items in each list were mixed to avoid obvious proximity of sentences with same gender ‘controllers’ or with ‘targets’ containing etymological in$ecting items.

!e testing was done in two phases. First, three older informants were tested for the four sentence types. Informants were presented the sentences orally and they were asked to ‘correct’ the sentences if they found them inappropriate. For the purposes of the study, it was assumed that if the informant corrected the ety-mological target in a sentence with a female ‘controller’, this was an indication of the vitality of the bound morpheme. From the "rst phase test we selected the 9 adjective pairs with the strongest vitality scores. !ese are indicated in bold in

!"#$%&''()

124 Alan N. Baxter

Table 1, in Section 6.14 In the second phase of testing the remaining speakers were evaluated in terms of the 9 adjectives identi"ed in the "rst phase, and in terms of the 10 nouns (in bold in the same table) with the strongest vitality scores registered when the words were assessed in isolation during the 2006 pilot study.

6. Vestigial etymological gender: derivational morphemes in MCP?

MCP is devoid of in$ectional morphology for NP number, does not have sub-ject-verb concord, and does not in$ect its verb for tense, aspect or modality. Furthermore, pronouns and most nouns and adjectives do not display gender. Where adjectives and nouns are concerned, MCP follows the tendency observed in the research literature on language acquisition and creole languages, adopting the masculine-gender form of the superstrate word as the unmarked (or default) form (Montrul 2004: 54, 78, 80; Holm 2000: 216). While neither Malay nor Hokkien, the primary and secondary substrate languages in Malacca, have in$ectional mor-phology, such in$ection is found in more distant potential substrates among the Indo-Aryan and Dravidian languages, a point which is further mentioned below.

Generally speaking, when it is essential to signal the sex of a nominal referent, MCP forms compounds with máchu ‘male’ and fémi ‘female’, as is also the case in Malay (Baxter 1988: 50):15

(1) a. báka máchu, báka fémi bovine male bovine female ‘bull, cow’ b. abó máchu, abó fémi grandparent male grandparent female ‘grandfather, grandmother’

Nevertheless, vestigial and possibly fossilized gender morphology has o%en been reported in the literature on MCP. !e earliest report, Rêgo (1942 [1998]: 63), compiled from "eldwork in the creole "shing community in Malacca in the early 1930s, noted that gender had been maintained in a good number of nouns, and in quite a few adjectives.16 !e examples that he cites are indeed pairs showing

14. !e adjectives mora ‘Indian’ and kĕriáda ‘adopted’ were not used as they cannot be focused within the NP.

15. !e MCP examples are given in the orthography used in Baxter (1988), and tonic stress is marked on all words with two or more syllables.

16. ‘Gender was preserved in many words: $lo, $la ‘son, daughter’; hómi, mulér ‘man, woman’; galo, galinha ‘rooster, hen’; consentidor, consentideira ‘abbettor’, etc. In other words, however,


a formal contrast that could be analyzed as etymological gender morphology. However, work on the dictionary of MCP (Baxter & de Silva 2004) has permitted a fuller identi"cation of the set of lexeme pairs that display a gender distinction.

Whereas in previous research on MCP (e.g. Baxter 1988: 49–50, 58–59; Tomás 2005: 139–40) this distinction was relegated to the lexeme itself, as an integral and unanalyzable part thereof, there may be a case for treating these distinctions as morphological. Firstly, it has been found that such pairs are somewhat more nu-merous than previously believed. Secondly, although there is variation in the use of certain words in these pairs symptomatic of ongoing loss of the gender distinc-tion, informants display a strong sense of the di&erence between the forms of these items, associate the pairs with their respective gender references and even point to the form of the word ending as conveying this di&erence.

In view of these considerations, MCP grammar may be said to display natural gender morphology (NGM) on items of [+hum] reference, such as kuzinyéra ‘fe-male cook’, kuzinyéru ‘male cook’. Is this morphology in$ectional or derivational? On the one hand, it resembles in$ection because it does not change the grammati-cal category of the word, nor does it derive a new word as such. It merely yields a di&erent form of the same lexeme, with the referent changing only in terms of the inherent property of natural gender. On the other hand, it shares with derivational morphology its limited scope as a local property of a restricted class of lexical items. In this sense, these pairs seem best treated as derivations. Certainly, mecha-nisms of etymological derivational morphology of some vitality have a signi"cant presence in the MCP lexicon.17

it disappeared, and only the more common form remains (…) !e adjectives exist more or less as in Portuguese with the only di&erence that gender and number have almost completely disappeared. !ere are, however, still quite a few adjectives that mainly retain gender: jogador, jogadera ‘gambler (m)/(f)’; minteroso, minterosa ‘deceitful (m)/(f)’; brigador, brigadera ‘"ghter (m)/(f)’, etc. (Rêgo 1942 [1998]: 63 [our translation]). Note, however, that jogadór, jogadéra and brigadór, brigadéra are mainly nouns in current MCP.

17. !e most frequently observed instances contain the etymological morphemes -du < P. -do, as in kebrádu ‘broken’ < kebrá ‘to break’; -sáng < P. -ção/-são, as in lembrasáng ‘memory, re-memberance’ < lembrá ‘remember’; -míntu < P. -mento as in padisemíntu ‘su&ering’ < padisé ‘to su&er’; and -(d)ór < P. -(d)or as in peskadór ‘"sherman’ < peská ‘to "sh’. Of these, the "rst, -du is proli"c, applying to virtually any causative verb base, and its derivations are readily ex-plained by informants as being the result of the perfective marker já plus the root verb. For example: kebradu ‘broken’ was suggested to be the result of já kebrá ‘(it) broke’. Of the others, both -sáng and -míntu apply to some 20 items each, and -(d)ór to over thirty. !ese derivational morphemes are to be found variously in the Indo-Portuguese creoles. For example, Dalgado (1998: 80, 89) registers the use of nominal derivations in Ceylon CP using the a#x -ção, and de-verbal adjectives with -do. Obviously, a closer inspection of accounts of the Indo-Portuguese creoles will reveal re$exes of these and other MCP etymological derivational morphemes. !e

!"#$%&''()

126 Alan N. Baxter

In MCP, NGM applies to at least 39 word pairs (i.e. some 78 words), of which the majority are the nouns shown in Table 1. !e nouns comprise kin terms (e.g. néta ‘granddaughter’, nétu ‘grandson’), certain other nouns concerning interpersonal relations, (e.g. biúba ‘widow’, biúbu ‘widower’), ethnicity (móra ‘Indian female’, móru ‘Indian male’), and endeavors (alkubitéra ‘procuress’, alkubitéru ‘pimp’), and a number of de-verbal agentive nouns (e.g. kuzinyéra ‘cook [f]’, kuzinyéru ‘cook [m]’). In addition, gender morphology is found in a diverse set of adjectives,23 also shown in Table 1, applying to human physical and character qualities (e.g. béla ‘old [f]’, bélu ‘old [m]’; prigasóza ‘lazy [f]’, prigasózu ‘lazy [m]’).

Looking to the o&shoots of MCP, we "nd certain similarities.24 !us, in Macau CP, in Table 2, based on Ferreira (1996), NGM applies to a set of [+hum] nouns

potential diachronic connection between the vestigial derivational morphology of these creoles is a worthy topic, and has special relevance to the types of input available in Portuguese Malacca. However, it would take us beyond the purpose of the present section.

18. Neither {bem} nor {fétu} have free independent distributions in MCP.

19. Here there is an additional linking item {-r-}. !e base, identi"ed by informants, is derived from MCP mídu ‘to fear’.

20. !is word was recognized by all informants, yet its masculine counterpart was recognized by only 6 informants.

21. !is word is synchronically denominal.

22. !e segment /er/ is not analyzable as an agentive separate from the segment /alkubit/, as the latter doesn’t occur elsewhere.

23. Some elderly informants in the 1980s claimed that there was more gender in$ection in the past. In this respect, it is interesting to note that both Rêgo (1942 [1998]), in the 1930s, and Knowlton (1964), in the 1960s, reported a gender distinction in the third person subject pro-nouns: éli (m), éla (f). !ere is no sign of this today.

24. Apart from the o&shoots of MCP, etymological gender morphemes are also to be found in some of the Indo-Portuguese creoles. However, in the northern varieties, for example Diu, Daman and Korlai, the loss of "nal unstressed vowels in words derived from Portuguese (cf. Dalgado 1998: 78, Clements 1996: 94) has for the most part eliminated traces of etymological gender morphology. !ere are some lexicalized exceptions, such as in Korlai CP irmãw ‘brother’ and irmã ‘sister’ (Clements 1996: 274–275). Cardoso (2009: 119, 263–264) reports that there is no etymological gender morphology as such, but there is a lexicalized distinction of natural gender in a set of human referent nouns. !e information available for the documented south-ern Indo-Portuguese creoles (Cannanore and Mahé [Schuchardt 1889], Mangalore [Schuchardt 1883], Cochin [Schuchardt 1882], Nagapatinam [Dalgado 1917]), which for the most part conserve word "nal unstressed vowels, is insu#cient to draw conclusions regarding the sta-tus of etymological gender in these varieties. In the case of Sri Lanka, Dalgado (1998: 81–83) reports that there is no grammatical gender on nouns or adjectives, yet notes that nouns with


Tabl

e 1.

Krist

ang:

noun

s and

adjec

tives

with

etym

olog

ical g

ende

r in$

ectio

n –

spea

ker r

ecog

nitio

n of

fem

inin

e ite

m (x

/15)

Nou

ns: N

on-d

eriv

edx/

15N

ouns

: Der

ived

x/15

Adje

ctiv

es: N

on-d

eriv

edx/

15Ad

ject

ives

: Der

ived

x/15

-a /

-u

de-v

erba

l age

ntiv

es-a

/ -u

de-a

djec

tival

1. n

éta ‘g

rand

-dau

ghte

r’15

-(d)é

r-a /

-(d)ó

r1.

béla

‘old

’ (F)

15-e

r-a /

-er-u

nétu

‘gra

nd-s

on’

1. a

rmad

éra

‘insti

gato

r’ (F

)4

b

élu (M

)1.

gro

séra

‘lech

erou

s, di

scou

rteou

s’(F)

14

2. d

isnéta

‘gr

eat-g

rand

-dau

ghte

r’9

a

rmad

ór (M

)2.

bem

féta18

‘han

dsom

e, be

autif

ul’ (

F)15

gr

osér

u (M

)

d

isnítu

‘gre

at-g

rand

-son

’2.

ase

ntid

éra

‘joke

r, ac

com

-pl

ice’ (

F)9

b

emfét

u (M

)-o

z-a

/ -oz

-u

3. p

ríma

‘cous

in’ (

F)8

a

sent

idór

(M)

3. b

uníta

‘bea

utifu

l, pre

tty’

(F)

62.

mid

uróz

a19 ar

ch ‘c

owar

dly,

fear

ful’ (

F)14

p

rímu

‘cous

in’ (

M)

3. b

rigad

éra

‘"gh

ter’

(F)

13

bun

ítu (M

)

mid

uróz

u (M

)4.

fíla

‘dau

ghte

r’ 15

b

rigad

ór (M

)4.

dód

a ‘cr

azy’

(F)

15

fílu

‘son

’4.

dib

inya

déra

‘for

tune

-telle

r’ (F

)9

d

ódu

(M)

de-n

omin

al

5. t

ía a

rch.

‘aun

t’8

d

ibin

yadó

r (M

)5.

gód

ra ‘f

at’ (F

)15

-oz-

a / -

oz-u

t

íu a

rch.

‘unc

le’5.

jug

adér

a ‘ga

mbl

er’ (

F)13

g

ódru

(M)

3. m

alisi

óza

‘lech

erou

s, lu

stful

’ (F

)11

6. n

ya ‘g

od-m

othe

r’14

ju

gadó

r (M

)6.

kaí

nya

arch

. ‘stin

gy’ (

F)10

m

alisi

ózu

(M)

n

yu ‘g

odfa

ther

’6.

kan

tadé

ra ‘s

inge

r’ (F

)6

k

aíny

u (M

)4.

min

taró

za ‘d

eceit

ful’

(F)

157.

sóg

ra ‘m

othe

r-in-

law’

15

kan

tadó

r (M

)7.

mór

a ‘In

dian

’ (F)

15

min

taró

zu (

M)

s

ógru

‘fat

her-i

n-law

’7.

kun

sintid

éra

‘abet

tor’

(F)

9

mór

u (M

)5.

prig

asóz

a ‘l

azy’

(F)

158.

min

ína

arch

. ‘’ba

by gi

rl’2

k

unsin

tidór

(M)

p

rigas

ózu

(M)

m

inín

u ar

ch. ‘b

aby b

oy’

8. t

apad

éra

‘con"

dant

’ (F)

2

!"#$%&''()

128 Alan N. Baxter

Tabl

e 1 (c

ontin

ued)

Nou

ns: N

on-d

eriv

edx/

15N

ouns

: Der

ived

x/15

Adje

ctiv

es: N

on-d

eriv

edx/

15Ad

ject

ives

: Der

ived

x/15

9. b

éla ‘o

ld w

oman

’15

t

apad

ór (M

)de

-ver

bal

b

élu ‘o

ld m

an’

-ér-a

/ -é

r-u10

-d-a

/ -d

-u9.

ko

mpa

nyér

a ‘co

mpa

nion

’ (F

)10

. biú

ba ‘w

idow

’2015

kom

pany

íru (M

)7.

kĕr

iáda

‘ado

pted

’ (F)

15

biú

bu ‘w

idow

er’

10.

kuzin

yéra

‘coo

k’ (F

)14

kĕ

riádu

(M

)11

. ful

ána

‘so-a

nd-s

o’ (F

)5

kuzin

yéru

(M)

f

ulán

u ‘so

-and

-so’

(M)

12. n

óiba

‘brid

e, gi

rlfrie

nd’

15de

-nom

inal

n

úibu

‘gro

om, b

oyfri

end’

-ád-

a / -

ád-u

13. f

ostér

a ‘st

rang

er, o

utsid

er’

(F)

911

. ch

umbu

áda21

‘laz

y per

son’

(F)

8

f

ostér

u ‘st

rang

er, o

utsid

er’

(M)

chum

buád

u (M

)

14. a

lkubi

téra22

‘pro

cure

ss’10

-er-a

/ -e

r-u

alku

bitér

u ‘pi

mp’

12.

pabi

séra

‘vul

gar,

foul

-m

outh

’ (F)

15

pa

bisé

ru (M

)


Tabl

e 2.

Maq

uista

(Mac

au C

reol

e Por

tugu

ese)

— n

ouns

and

adjec

tives

cont

ainin

g ety

mol

ogica

l gen

der i

n$ec

tion

Nou

nsAd

ject

ives

Non

-der

ived

Der

ived

(de-

verb

al an

d de

-adj

ectiv

al

agen

tives

)N

on-d

eriv

edD

eriv

ed(d

e-ad

ject

ival

, de-

verb

al)

de-verba

lde-adjectiv

al

-u/-a

1.

pórc

a ‘so

w’

pó

rco ‘

pig’

2.

sógr

a ‘m

othe

r-in-

law’

sógr

o ‘fa

ther

-in-la

w’3.

nó

iva ‘b

ride’

nói

vo ‘g

room

’

-áng

/-á4.

la

drá

‘thief

’ (f)

l

adrá

ng (m

)5.

irm

á ‘si

ster’

irmán

g ‘br

othe

r’6.

an

á ‘dw

arf’

(f)

anán

g (m

)

-ang

/-oa

7.

lióa

‘lion

ess’

l

iáng

‘lion

’8.

pa

vóa

‘pea

hen’

pavá

ng ‘p

eaco

ck’

9.

patró

a ‘b

oss’

(f)

patrá

ng (m

)

-ór/-

óra

1.

cant

óra

‘sing

er’(f

)

cant

ór (m

)2.

pi

ntór

a ‘p

ainte

r’ (f

)

pint

ór (m

)-(d

)ór/-

(d)é

ra3.

ab

usad

éra

‘bol

d, in

solen

t’ (f

)

abus

adór

(m)

4.

vare

déra

‘swe

eper

’ (f)

va

redó

r (m

)5.

ve

nded

éra

‘selle

r’(f)

ve

nded

ór (m

)6.

na

mor

adér

a ‘$

irt’ (

f)

nam

orad

ór (m

)7.

ga

stadé

ra ‘s

pend

er, o

ne w

ho w

aste

s m

oney

’

gasta

dór (

m)

-áng

/-óna

8.

man

dóna

‘bos

sy p

erso

n’ (f

)

man

dáng

(m)

9.

vilón

a ‘vi

llain

’(f)

vi

láng

(m)

10.

piliz

óna

‘con$

ictiv

e per

son’

(f)

pi

lizán

g (m

)11

. sa

bich

óna

‘kno

w-all

’ (f)

sa

bich

áng (

m)

de-adjectiv

al-Ø

/-a12

. im

postó

ra ‘i

mpo

stor’(

f)

impo

stór (

m)

-u/-a

1.

chist

ósa

‘plea

sant

, nice

(per

son)

’ (f)

ch

istôs

o (m

)2.

dó

da ‘c

razy

’ (f)

dô

do (m

)3.

m

óna

‘stup

id’(f

)

môn

o (m

)4.

tó

la ‘f

oolis

h, n

aïve’

(f)

tô

lo (m

)5.

gó

rda

‘fat’(

f)

gôrd

o (m

)6.

bó

ba ‘f

oolis

h, n

aïve’(

f)

bôbo

-Ø/-a

7.

istop

óra

‘dope

y’ (f

)

istop

ôr (m

)

-áng

/-óna

8.

pim

póna

‘row

dy’ (

f)

pim

páng

(m)

-ang

/-óna

1.

sobe

rbón

a ‘(e

xces

sively

) arr

ogan

t’ (f

)

sobe

rbán

g (m

)2.

fei

óna

‘very

ugl

y’ (f

)

feián

g (m

)

de-verba

l-a

ng/-ó

na3.

ch

orón

a ‘(r

eadi

ly) co

mpl

ainin

g’ (f

)

chor

áng (

m)

!"#$%&''()

130 Alan N. Baxter

and adjectives with [+hum] referents, comparable to its distribution in MCP.25 Here too, the nouns include kin terms (e.g. sógra ‘mother-in-law’, sógro ‘father-in-law’), interpersonal relations, (e.g. nóiva ‘bride’, nóivo ‘widower’), endeavors (ladrá ‘thief (f)’, ladráng ‘thief (m)’), and de-verbal and de-nominal agentive nouns (cantóra ‘singer (f)’, cantór ‘singer (m)’). In addition, gender morphology is found in a set of adjectives relating to human physical and character traits (górda ‘fat (F)’, górdo ‘fat (m)’; bóba ‘stupid (f)’, bóbo ‘stupid (m)’). However, etymological gender is also found in some [−hum, +animate] items as well: gálu, galínya ‘rooster, hen’; pórco, pórca ‘pig, sow’, paváng, pavóa ‘peacock, peahen’ (Ferreira 1996: 231–234). In MCP there is only one animate pair that could so qualify: gálu, galínya ‘rooster, hen’, as is also the case in the available data on Tugu (Schuchardt 1891: 133).

In Tugu CP, NGM applies to the small number of [+hum] nouns shown in (2).26 However, no evidence was located indicating NGM in adjectives.

(2) a. intiádu, intiáda ‘stepson’, ‘stepdaughter’ b. dónu, dóna ‘owner (m)’, ‘owner (f)’ c. tíu, tía ‘uncle’, ‘aunt’ d. fílu, fíla ‘son’, ‘daughter’ e. subrínyu, subrínya ‘nephew’, ‘niece’ f. kunyádu, kunyáda ‘brother-in-law’, ‘sister-in-law’ g. $ládu, $láda ‘Godson’, ‘Goddaughter’ h. sirbidór, sirbidéra ‘servant (m)’, ‘servant (f)’

!e study of the scope of NGM can also be extended beyond the Asian creoles. Given the limitations of space, we do not intend to make extensive comparisons with other Portuguese-based creoles, but simply register the existence of similari-ties with the distribution noted in MCP. !us, we "nd similarities in the treatment of natural gender in contemporary Cape Verdean, where the [+hum] feature, in particular, yet also [−human, +animate], are the prime determiners of a noun’s capacity to carry gender morphology. Baptista (2002: 42–43) presents examples of

the agentive a#x -dor form the feminine in -deira, as in mercador, mercadeira ‘merchant/seller (male, female)’, yet also in -ora, as in peccadora ‘sinner (female)’.

25. It is important to note that this table is meant only to provide a sample. It is not based on a study addressing gender morphology.

26. !e examples are from (Schuchardt 1891: 115–116). !e original Dutch orthography has been replaced by the MCP orthography used in Baxter (1988) and tonic stress is indicated on words of two or more syllables.


such [+hum] referent nouns that fall into three classes: kinship terms, activities performed by humans, and aristocratic titles, as in (3):27

(3) a. fídju, fídja ‘son, daughter’28

b. nóibu, nóiba ‘groom, bride’ c. ladrón, ladróna ‘thief (m), thief (f)’ d. trabadjadór, trabadjadéra ‘hard worker (m), hard worker (f)’ e. papiadór, papiadéra ‘someone who talks a lot (m), (f)’ f. dúki, dukéza ‘duke, duchess’ g. príspi, priséza ‘prince, princess’ h. kóndi, kondésa ‘count, countess’

While most of the items listed in Table 1 are fairly stable, some display variation.29 In these cases, the feminine gender distinction is only variably applied, and the distinction is lost when the former masculine member of the pair is applied to both masculine and feminine referents. Similar facts are found in contemporary Cape Verde Creole (Baptista 2002: 66–67). Variation in NGM in MCP allows us to witness the synchronic dimension of this diachronic process, whereby some distinctions are being lost. !e process proceeds at di&erent rates for di&erent speakers, creating three coexistent, overlapping systems, as in (4). !e endpoint of the process is the invariant form applicable to both sexes.

(4) ‘gambler’ jugadéra, jugadór → (jugadéra), jugadór → jugadór (f) (m) (f) (f) + (m) (f) + (m)

In searching for explanations for the retention of bound morphology for natural gender in MCP, in the face of a predominantly non-in$ecting NP, several factors seem relevant. Communication pragmatics and word frequencies may be a sig-ni"cant explanatory factor, as most of the nouns in Table 1 are reasonably central to basic social relations. On the other hand, the greater majority of these items

27. Certain of the Spanish-based creoles display certain similarities with the facts in the Portuguese-based Creoles. In the Asian region, in Ternateño (Patrick Steinkrüger p. c.), in Zamboanga Chabacano (Forman 1972: 102, 104, 123–124), and also in Ermiteño and Caviteño (Mauro Fernández, p. c.), etymological in$ection expressing natural gender restricted to [+hum] referents is also found with nouns and some adjectives. Some similar, yet minor traces of this are also found in Papiamento (Munteanu 1996: 266) and Palenquero (Schwegler & Green 2007: 294; Schwegler p. c.).

28. !e spelling in all the Cape Verdean examples has been modi"ed to show tonic stress.

29. Some other cases, such as tíu/tía have become obsolete, having been replaced with English-derived terms ángkĕl/énti. Yet others, such as minínu/minína, are known only to a restricted number of older speakers.

!"#$%&''()

132 Alan N. Baxter

involve gender marking in conjunction with other bound morphology. !us, an additional motive for the retention of gender marking in these cases might be their morphological complexity and phonological weight, a saliency factor. In the latter regard, an additional factor that permeates all four groups in Table 1 is that eighteen of these word pairs involve metaphonic gender marking, which might be considered an extension of a tendency already present in Portuguese morphopho-nology. !ey are essentially double-marked for gender by means of a regressive vowel height assimilation triggered by the word-"nal vowel of the etymological in$ection and applying to the tonic vowels /e/ and /o/:30

(5) ["s#gra] ["sogru] parent-in-law.f parent-in-law.m ‘mother-in-law’ ‘father-in-law’

!e degree of raising varies somewhat according to speaker. However, informants draw attention to the di&erent quality of the tonic vowel. In some pairs, this pro-cess even raises [e] to [i] (kompanyíru ‘companion’), and [o] to [u] (núibu ‘groom, boyfriend’). !e redundant marking involved in metaphonic gender pairs might be viewed as enhancing retention of natural gender morphology.

Finally, there may be a case for positing an in$uence of Indo-Aryan and Dravidian languages on the development of MCP, stemming from Indian forbears of the community. Noun gender morphology in these language families displays orientation by natural gender. In the Indo-Aryan languages, the system of gender (i.e. gender + natural gender) is similar to that of Portuguese. Masica (1991: 218) notes that northern Indo-Aryan languages, ‘due to a social need to distinguish female members’, commonly have gender morphology to distinguish male and female members of occupational groups and also kin. However, it is the Dravidian family that seems more pertinent. For example, in Tamil, nouns fall into three morphological gender classes according to whether they denote a male human (or deity), a female human (or deity), or other referents. !is distinction is shared by several other Dravidian languages, including Kannada (Corbett 1991: 8–11, 2006: 126; Krishnamurti 2003: 205–206), but also Sinhala, an Indo-Aryan lan-guage of Sri Lanka (Gair & Lust 1998: 8–9). !e Dravidian languages could have provided the closest model to what is found in MCP.

30. !e process may be summarized as V.mid → V.midαheight/__C Vαheight##. However, we are not aware of metaphony a&ecting words in Portuguese that incorporate derivational morphemes in conjunction with gender in$ection.


7. Does MCP have vestiges of contextual gender in.ection?

It seems logical to ask whether there are vestiges of contextual gender in$ection in MCP. !at is, is there a morphosyntactic dimension to NGM? Indeed, Hancock (1969: 40) noted that ‘vestiges of noun-adjective concord are to be found with a few adjectives’. Such sequences appear to constitute two types.

Firstly, there are sequences within NPs where the adjective may or may not bear an etymological gender morpheme in apparent consonance with the inher-ent gender of the head noun. !is is shown in example (6a), where the head has feminine gender reference. In contrast, in (6b) the head has masculine reference and only admits the etymological masculine gender adjective:

(6) a. mulé gódra /gódru, mulé mintaróza/ mintarózu woman fat.f fat.um, woman lying.f lying.um ‘fat woman’, ‘deceitful woman’ b. ómi *gódra/ gódru man *fat.f/ fat.um ‘fat man’

!is variable pattern of gender matching is also to be found in Cape Verde Creole data (Baptista 2002: 66–67). In this case, a feminine animate noun may co-occur either with the etymological feminine or masculine form of an adjective, as in example (7a). However, a masculine animate noun is only modi"ed by an etymo-logically masculine adjective, as in example (7b):

(7) a. un mudjér bránka/ bránku art woman white.f white.um ‘a white woman’ b. un ómi *bránka/ bránku art man *white.f white.m ‘a white man’

!e second type of MCP N+ADJ matching pattern that might be considered as involving a vestige of contextual in$ection is shown in examples (8–11). In this type of sequence, the substitution of the default (non-gender marked) adjective is blocked whenever the head has feminine natural gender.

(8) a. fíla kĕriáda/ *kĕriádu daughter adopted.f *adopted.m ‘adopted daughter’ b. fílu kĕriádu son adopted.m ‘adopted son’

!"#$%&''()

134 Alan N. Baxter

(9) a. mulé móra/ *móru woman Indian.f *Indian.m ‘Indian woman’ b. ómi móru man Indian.m ‘Indian male’

(10) a. fíla mína béla/ *bélu daughter virgin old.f *old.m ‘spinster’ b. mansébu bélu batchelor old.m ‘elderly batchelor’

(11) a. ńgua fémi bemféta/ *bemfétu one female handsome.f *handsome.m ‘a pretty young lady’ b. ńgua máchu bemfétu one male handsome.m ‘a handsome young man’

!us MCP displays, albeit marginally, two apparent gender matching sequences within the nominal domain. We will simply call these variable matching and strict matching in order not to invoke any morphosyntactic notion of agreement and thereby convey the mistaken idea that we are arguing for the existence of agree-ment in MCP. In terms of descriptive e#ciency, it would be excessive to motivate a gender agreement rule on the basis of relatively minor facts of distribution. A simpler solution would be to treat such cases as compounds, "xed in examples (8–11) and variable, in example (6), where a change to the default form of the adjective is underway.

!e instances considered so far involve contiguous head-modi"er sequences. However, these sequences do not meet two of the traditional criteria for classi"ca-tion as compounds: non-separability and structural integrity (Arono& & Fudeman 2005: 37–38; Matthews 1991: 99). !e N+ADJ sequences in the above MCP ex-amples fail the criterion of non-separability because, in each instance, all our in-formants accept modi"cation of the adjectives:

(12) ńgua fémi bomóng bemféta/ *bemfétu one female good-good handsome.f *handsome.m ‘A very attractive young lady’

Furthermore, for all the speakers considered here, the criterion of structural in-tegrity is not met. In MCP a syntactic process can be applied to the adjective in an


NP, shi%ing it over the noun and placing it prior to the DET, demonstrating that the adjective is separable from the noun. !is is the focusing rule referred to in Section 5 which is shown here in example (13). It is common in NPs which func-tion as emphatic recasts or as independent exclamations.

(13) bomóng bemféta/ *bemfétu ńgua fémi good-good handsome.f *handsome.m one female ‘A pretty young lady’

Here the adjective bemféta,bemfétu is le%-dislocated for focus, and placed phrase-initially (Baxter 1988: 193–194).31 Yet, for the speakers of the older set and for most of the younger set, le%-dislocated bemféta and bemfétu are restricted respec-tively to feminine and masculine nominal heads. For the older set of speakers considered, the rule applies with all adjectives that display etymological gender morphology. !is fact undermines the compound status of N+ADJ and is sugges-tive of a minor vestige of intra-phrasal gender concord.

Table 3 shows the acceptance scores of 15 speakers for the 9 key NGM ad-jectives in the three structures: N+ ADJ, N+ ADJ + ADJ, and le%-dislocated (ADV)+ADJ +N:32

In the available data for Tugu CP, there is no evidence of intra-phrasal gender matching. However, in current moribund Macau CP, as is to be expected through ongoing contact with Portuguese, there is a great deal of variable gender match-ing. Ferreira (1996: 233–234), in his account of conservative Macau CP,33 notes that there is both absence of matching,34 where adjectives have only the unmarked form, as well as variable matching. However, he gives no indication of whether any adjectives enter into invariant matching. !is matter is open for further research.

31. !is dislocation rule is also found in Tugu CP and, as Schuchardt (1891: 113, 254–255) notes, also in Ceylon CP: see examples in Dalgado (1998: 147, 148, 150, 158, 165, 167) and in Mello (1998: 151, 153, 160, 163, 167 ). One example of this structure is also recorded in Mangalore CP (Schuchardt 1883: 888).

32. Note that the feminine reference nouns employed in the evaluation exercise did not contain etymological gender morphemes. !e items fémi ‘female’ and mulé ‘woman’ were used as the heads. No attempt was made to evaluate whether the presence of a noun with an etymological gender morpheme exerted any e&ect on the selection of the adjective form. !is would be a topic for further research.

33. Ferreira’s account, compiled in the early 1970s, was based on the language of elderly conser-vative speakers and his own speech.

34. Ferreira actually uses the term concordância ‘agreement’.

!"#$%&''()

136 Alan N. Baxter

Table 3. NP with NGM feminine adjective and feminine reference nounNGM ADJ N(F)+ADJ N(F)+ADV+ADJ (ADV)+ADJ+Nbéla33 11 13 13bemféta 15 15 15doda 15 15 15gódra34 13 14 15groséra 10 11 12midaróza 11 12 13prigasóza 15 15 15mintaróza 13 14 15malisióza 13 11 11

Another area of interest here involves the insertion of NGM adjectives and nouns into predicative constructions, as in examples such as (14) and (15) below. While these predicates admit some variation in the use of the male or female etymo-logical items of certain NGM pairs, informants preferably match inherently fe-male gender subjects and inherently female predicatives displaying etymological gender. With the 10 NGM nouns preselected for their prominence in the recog-nition test, shown in bold in Table 1, the matching is categorical.37 !is type of predicative is shown in the following examples, where both the subject and the predicate contain an NGM item with etymological gender. In (14) the predicate contains an adjective and in (15) a noun.38

(14) éli sa fíla bomóng mintaróza/ *mintarózu 3s poss o&spring.f good-good deceitful.f *deceitful.m ‘Her/his daughter is very deceitful’.

35. !e low acceptance of béla, in spite of the fact that it registers a strong score when evaluated as an isolated word, is owed to the fact that there was disagreement among speakers as to the use of the item as an adjective. Some younger speakers preferred idádi ‘age’.

36. While all informants preferred gódra in the N+ADJ and N+ADV+ADJ structures, three informants (20%) stated that it was acceptable to use gódru.

37. As in the case of examples (5), (7–11), examples (12–15) are based on the observation of natural data and the subsequent evaluation of the grammaticality of test sentences. Testing con-ducted recently, in April 2009, con"rmed a tendency observed in the earlier pilot study: there is an overall preference for the F agreement amongst all speakers, which is stronger among older speakers, and categorical for certain NGM items.

38. MCP does not have a copular verb (Baxter 1988: 182–183).


(15) aké béla ńgua grándi pabiséra/ *pabiséru dem old.f one big vulgar.f *vulgar.m ‘!at old lady is a real foul-mouth’.

Item (15) brings to mind such politically incorrect English examples as the hostess is an authoress. Certainly, in English and in MCP, given the small number of nouns with gender a#xes, it seems illogical to treat these as anything but straightforward instances of pragmatic and semantic selection of individual lexical items that have the inherent characteristic of gender. However, where adjectives are concerned in MCP, in spite of the very limited and speci"c nature of the class, it seems descrip-tively and theoretically inappropriate to list adjectives individually as having in-herent gender features. !is does not appear to be a simple case of lexical selection. Rather, these do seem to represent a restricted type of dependent marking (Booij 2007: 104), albeit a restricted vestigial instance thereof. In examples such as (14), there is a dependency relation between the head, the subject, and a dependent (the predicate adjective), involving the matching of an inherent quality of the head that is marked on the dependent.39

In comparison, in the available data on Tugu/Batavia CP, while it may be as-sumed that the NGM nominal pairs in (2) could be inserted into equational predicates (and this could be handled by treating them as individual lexical items), there is no evidence of intra-phrasal matching involving adjectives. On the other hand, in Macau CP, there is some evidence of variable interphrasal agreement in-volving predicate nominals, which may be attributed to a context of continued con-tact with Portuguese. For reasons of space limitations, we will not enter into details.

8. +e status of gender morphology in MCP

In the previous sections, we saw that MCP has, in addition to etymological derivation-al morphology, some inherent gender morphology and limited traces of contextual

39. A similar situation appears in Zamboanga Chabacano which has evidence of matching of natural gender [+hum] referents represented by items containing etymological in$ectional morphemes. !is is found in intraphrasal instances, such as in boníta prinsésa (beautiful.f + princess) ‘beautiful princess’ (Forman 1972: 104), and also in interphrasal instances, between subject and predicate adjective, as in (i):

(i) komó prinsésa siémpre boníta (ibid.:41–42) since princess always beautiful.f ‘Since princesses are always beautiful’

Mauro Fernández (p. c.) con"rms that natural gender matching in Chabacano involving items containing etymological in$ectional morphemes can occur over several clauses.

!"#$%&''()

138 Alan N. Baxter

gender morphology both at intraphrasal and interphrasal levels. In one sense, this pro"le matches what is widely found in creoles that are claimed to have in$ection: a greater degree of inherent in$ection than contextual in$ection. !e fact that it has minor traces of contextual dependent marking suggests that second language acqui-sition (SLA) did reach a level of intraphrasal (and possibly incipient interphrasal) morphological development involving the nominal domain. However, the retention of inherent in$ection reported so far in creoles seems to involve more the category of number. Indeed, Bakker (2003) noted that in the sample of pidgins that he surveyed, the retention of inherent number in$ection was more common and that of gender less so. Yet, where the NP is concerned, MCP does not retain in$ectional number. Indeed, we "nd only a couple of words that have entered the language with their plural morphemes fused and unanalysed: máris ‘wave, waves’, bías ‘vein, artery’. Why, then, should MCP retain a re$ex of Portuguese in$ectional gender?

!ere may be good reasons for this associated with both the con"guration of number and gender within grammar and the acquisition process itself. In the "rst language acquisition (FLA) area, some researchers in language processing, working with L2 Spanish, have proposed that number attribution results from the building of phrase structure, whereas gender is already present in the lemma representation (Domínguez et al. 1999; Franceschina 2005). Indeed, Franceschina (2005: 101–104) discusses a body of research that points towards a clear di&erence between number and gender, suggesting that the gender a#x displays a stronger connection with the noun stem than does the number a#x. !is is consonant with a view of generative syntax whereby number originates in the head of a NumP, yet gender is an inherent feature of the lexical entry of the noun (Kester 1996, White et al. 2004, Franceschina 2005). From the perspective of Processability !eory, the fact that attribution of number to the noun would require movement would have to be seen as a more complex step than the category procedure that assigns the gram-matical category of the head, triggering subsequent phrase building (Pienemann, 2005: 7–14, Plag 2008). It would need to be seen as part of phrase building.

Further support for the sequencing of gender prior to number, or its struc-tural positioning in an earlier location in an incremental procedure, may be drawn from the "ndings of acquisition research, where it has been widely observed that in both FLA and bilingual language acquisition (BLA) of Spanish, number agree-ment develops later than gender agreement (Montrul 2004: 54, 84; Franceschina 2005: 113). In contrast, in the SLA area, the picture is not so clear, for work on the acquisition of gender and number in Spanish relates mainly to the implementation of agreement in intermediate and advanced level L2 speakers.40

40. In intermediate and advanced Spanish L2 data it has been found that number agreement outpaces gender agreement (Montrul 2004: 81, White et al. 2004). Nevertheless, it again needs to


Nevertheless, whether in FLA, BLA, or SLA, it is assumed that the structure that language acquirers assemble involving gender and number is the same. If gender is earlier in place in the N than number, as a required root feature of the N, whereas number requires installation via movement, it seems logical to postulate a develop-mental phase in which gender would be (momentarily) more visible than number. As such, the situation in MCP could indicate that SLA input to MCP was at an initial stage of NP development where gender and number are concerned, and N move-ment for number had not been installed. Furthermore, it can be assumed that the typologies of the languages in contact could have important rami"cations for this.

!ere remains the question of why the etymological gender morphology re-tained in MCP should involve natural gender referents only. Evidence from FLA and SLA research on the acquisition of Spanish suggest that of the three factors potentially involved in facilitating the development of gender morphology — mor-phophonological, morphosyntactic, and semantic (natural gender) clues avail-able to the learner, the "rst two are initially more important (Montrul 2004: 54, 79; Franceschina 2005: 109). Broadly speaking, the research literature does not claim a signi"cant role for natural gender in the acquisition of gender morphol-ogy, although certain studies of L2 Spanish (Anderson 1984, Finnemann 1992, Fernández-García 1999, Alarcón 2004) have shown that natural gender is impor-tant in early stage L2 acquisition. Furthermore, SLA research conducted from a processing perspective (Irmen & Knoll 1999) also shows the precedence of nat-ural semantic clues over morphosyntactic clues in their research on speakers of Finnish (a language with no grammatical gender) acquiring pronominal gender in German. !us, in the case of MCP, it seems plausible that the substrate speakers of the L2 Portuguese input to MCP found the morphosyntactic clues in Portuguese less noticeable than the natural semantic clues. While Plag (2008: 125) has pro-posed that ‘the semantic value expressed by in$ectional a#xes does not seem to determine their survival’, it could be that in certain language contact interfaces in-volving natural acquisition it may be easier to recognize/acquire gender in$ection

be stressed that the L1s in most of this research have number morphology. On the other hand, recent research on the acquisition of Portuguese gender and number morphemes by speakers of Cantonese shows that the presence of the appropriate gender morpheme favours the presence of the appropriate number morpheme on the same word (Godinho 2005: 345; Matos 2008: 76). Lucchesi (2000: 241) reports similar results in his study of the generational acquisition of Portuguese gender agreement by an Afro-Brazilian speech community in which a degree of cre-olization occurred in the 19th century. Work on conservative rural Afro-Brazilian Portuguese, in communities with histories of language contact, found variation in gender marking to be less prominent than variation in number marking (Baxter 1995).

!"#$%&''()

140 Alan N. Baxter

when this relates to a sex distinction in animates, especially of humans.41 In this regard, certain Indo-Aryan and Dravidian languages may have provided a relevant interface with natural gender morphology systems at an early stage in the develop-ment of MCP.42

In view of the above observations on the speci"c format of MCP inherent natural gender marking, it is proposed that the current status of this morphology in the language may be viewed from three perspectives, as resulting from:

a. SLA input to L1 acquisition, with SLA data from various sources involving overlapping developmental stages; or

b. SLA of morphological material from Portuguese by creole speakers in a dia-chronic diglossic situation interfacing Portuguese and a continuum of creole lects; or

c. Ongoing erosion of gender in$ection as a result of bilingualism in Malay and English.

In all three perspectives, acquisition and loss of bound morphology in MCP could be viewed as involving the same prime conditioning factors: internal structuring restrictions, pragmatic, semantic and morphophonological saliency, and transfer e&ects from substrate and adstrate languages. While point (c) accounts for what is happening currently in MCP among younger speakers, the dating of gender in$ection in MCP depends on determining the relative importance of (a) and (b). !e following section explores historical evidence with the aim of determining the chronology of diglossia among MCP speakers.

9. Did diglossia cease when the Dutch conquered Malacca?

In Bickerton’s (1988, 1999) terms, bound morphology, especially in$ectional mor-phology, was viewed as a potential early casualty in the dilution process involved in creolization, whereby superstrate morphemes are not acquired in the creole.

41. Lucchesi (2000) observes that variation in gender marking in Afro-Brazilian Portuguese af-fected the syntactic dimension more than the in$ection on the NP head, gender in$ection being favoured on [+hum] nominal heads.

42. !e in$uence of Indian substrates is felt elsewhere in the grammar of MCP, in dative-accu-sative and genitive marking (Baxter 1988: 93, 167–9). However, these features in MCP involve stand-alone morphemes in contrast to their bound counterparts in Dravidian and Indo-Aryan. Nevertheless, as Mauro Fernández (p. c.) points out, it does not seem necessary for the sub-strate languages to have a morphological pattern of natural gender in order for this feature to be acquired from another language: Tagalog does not have this feature, yet it has developed in Chabacano.


Yet, particular sociolinguistic contexts, and the contact between languages of dif-ferent morphosyntactic typologies, may lead to di&erent degrees of retention of morphology. !us, the extent of in$ectional verbal morphology in the CPs of Daman and Diu, in contrast with its absence in MCP, would be found in the di&er-ent sociolinguistic histories and di&erent language typologies involved in contact. Daman CP had a long presence of diglossia, as Daman was a Portuguese colony until 1961, and involved contact between Portuguese and Gujarati, both languages rich in bound morphology and especially in in$ection.43 So, what was the situa-tion in Malacca? Hancock (1975: 25) suggested that a diglossic situation had not existed in Malacca ‘for 300 years or more’. Did diglossia end with the Dutch take-over? It will be seen that this was probably not the case.

As yet, there is no information on the form of MCP earlier than 1830, and there is no linguistic data earlier than 1890 (Baxter 1996: 307–308).44 !ere is a good set of data from Tugu CP from the late 19thC (Schuchardt 1891) but it con-tains no direct evidence of contextual matching of items with etymological gen-der morphemes, and there is no evidence of etymological gender in$ection pairs among the adjectives. Nevertheless, while Tugu CP and MCP are quite similar in many other respects of their grammars, it must be remembered that, in 1891, 248 years had passed since 1642. !e di&erences between MCP and Tugu CP are linked to the di&erent sociolinguistic settings of these two languages since that time.

While we cannot answer diachronic issues in the absence of earlier linguistic data, it is possible to gain some insight into the sociolinguistics of MCP during the Dutch period via demographic and maritime trade data. On the eve of the Dutch takeover there were 250 European Portuguese casados with indigenous wives, a group of local converts numbering some 7400, and a group of some 2000 slaves (Maxwell 1911: 4). !ere would also have been a $oating Portuguese population in-volved in trade and military endeavours (MacGregor 1955: 6). Fernando (2004: 166, 175) notes that Dutch documentation in 1680 shows that the ‘Portuguese’ com-munity, although much reduced in size, retained (in 1680) the following pro"le: European (40), Mixed (61), Black (537), Slaves (529).45 Furthermore, records of

43. Korlai CP (Clements 1996) is rich in verbal morphology. Although the o#cial Portuguese presence in Chaul ended in 1740, Clements notes that there was subsequent contact with speak-ers of CP from other communities in India, and a presence of Portuguese-speaking priests in Korlai until 1964.

44. Two early documents in MCP were located in the Schuchardt archives, one representing a conversation in language very similar to that of today, and the other written by a missionary, part of a translation of the New Testament. !e latter text shows considerable in$uence from Metropolitan Portuguese.

45. !e terms Mixed and Black are from the Dutch classi"cation.

!"#$%&''()

142 Alan N. Baxter

shipping movements at the port of Malacca show that between 1682 and 1742, 78 ships were captained by Portuguese (not Mixed or Black) based in Malacca and for the period 1761–1792, 203 ships were captained by Portuguese (not Mixed or Black) also based in Malacca (Fernando 2004: 176, 178). While these shipping "g-ures cannot give a clear notion of the actual number of Portuguese involved (since many of the ships would probably have been captained by the same persons), they do suggest that there was still a presence of Portuguese of European descent.

Another relevant point is that Malacca continued to be linked to trading ports where the Portuguese were based. !us, in the period 1682–1742 (Fernando 2004: 177), Malacca was visited by ships captained by Portuguese who were based in Goa (56 ships), Macau (97 ships), Madras (62 ships), and in smaller numbers from additional locations where it is known that ‘Portuguese’ communities re-sided (Calicut, Mangalore, Nagapatnam, St !ome [de Meliapor]).46 For the pe-riod 1761–1792 (Fernando 2004: 179), the port was visited by ships captained by Portuguese based in Macau (828 ships), Bombay (19 ships), Goa (15 ships), Madras (45 ships) and Malabar (33 ships), and in smaller numbers from other locations where ‘Portuguese’ communities resided (Coromandel, Nagapatnam, Tranquebar). Furthermore, in the period 1682–1742 (Fernando 2004: 181), ships captained by Black Portuguese, originated from Macau (98 ships), Goa (58 ships), Madras (58 ships), and in smaller numbers from other Indian ports where ‘Portuguese’ com-munities resided (Mangalore, Nagapatnam, St !ome [de Meliapor]). Finally, in the period 1761–1792 (Fernando 2004: 183), ships captained by Luso-Asian per-sons classi"ed as Mixed originated from Macau (500 ships), Bombay (58 ships), Goa (72 ships), Malabar (128 ships) and Madras (140 ships), including smaller numbers from other known locations of ‘Portuguese’ communities (Coromandel, Nagapatnam, Tranquebar).47

Some further evidence that lends support to the notion of a prolonged diglos-sia comes from the areas of public administration, religion and education, and from the testimonies of elderly speakers in 1980. Firstly, late 18th and early 19th century documentation from the Dutch and early British administrations con-tains some items in metropolitan Portuguese. !e renewal of an o#cial presence of Portuguese-speaking priests (mainly European-born) from Goa (and Macau) dates from the "rst decade of the 18thC (Baxter 1988: 8). !is constant presence continued until the late 20th century, and until WWII, European Portuguese priests continued to use both Portuguese and MCP in the church. Some of these

46. All the names of ‘places’ are those identi"ed by Fernando (2004) in the Dutch records. !e references here also include shipping from the coastal region of Malabar without mention of particular ports.

47. !e references include shipping from the Coromandel region without speci"cation of ports.


priest-register materials were published in Rêgo (1942 [1998]) and show a certain in$uence from Metropolitan Portuguese.

Secondly, some Portuguese schools operated in Malacca in the 19th centu-ry, especially in the "rst half of the century under the auspices of the Portuguese Mission, the London Missionary Society, and the Malacca Free School (Baxter 1996: 307). In 1832, the London Missionary Society ran "ve Portuguese schools catering for 80 boys, 32 girls, and 10 female adults, and evidently both Portuguese and Creole Portuguese were used in these schools (Baxter 1996: 308). Indeed, one missionary requested of her London superiors ‘A few Indo-Portuguese Testaments (+ Bibles if rendered into that dialect) would be very acceptable — remember that the Malacca Portuguese do not speak the Madras Indo Portuguese — It is that of Ceylon which approaches nearest to our [emphasis in original] local dia-lect’ (Garling 1830). Some of the records of the 19th century Portuguese Mission Schools are written in Portuguese, yet not compiled by a priest, so clearly some Portuguese Eurasians had knowledge of Portuguese (Baxter 1988: 10). In the early 20th century, a Convent school with some nuns from Macau received Portuguese Eurasian girls, and also taught some Portuguese (Baxter 1988: 10, 1996: 307).

Finally, further information that would support the past existence of a lectal spread comes from the testimonies of elderly MCP speakers interviewed in 1980. !ey point to an interesting detail: that the MCP of the better-o& sector of the community that formerly resided in the Tranquerah area of Malacca town (the area of residence of the traders) was di&erent, more fúndu ‘deep’, than the MCP of the "shing community.

In sum, there is historical evidence that could implicate a degree of continued diglossia, and at least a wider Creole lectal grid. Such factors might have helped to maintain NGM in MCP, or could even have introduced it in the centuries af-ter the Dutch takeover.48 !e complexity of the sociolinguistic history of MCP and the absence of early linguistic data render the dating of the introduction of NGM highly problematic. While inherent gender morphology may be attributed to early-stage SLA processes, it seems logical that there would be a role for SLA input, and early stage SLA structuring processes, at more than one time: at initial creole formation stages and also in an ongoing diglossic setting where the L1 cre-ole speakers may acquire L2 Portuguese.

48. Macau CP and Cape Verdean may be instructive in this respect, as their ongoing contact with Portuguese implicates NGM on a wider range of items. In the earliest linguistic description of Cape Verde CP, Brito (1887 [1967]: 349), a native speaker of the creole, did not report etymo-logical gender morphology [+hum] items. Baptista (2002: 66) notes that acrolectal varieties of Cape Verdean are more likely to display some gender agreement. !us, NGM now observed in that creole may result from contact with Portuguese since the 19th century.

!"#$%&''()

144 Alan N. Baxter

10. Conclusion

MCP has etymological gender morphology in a consistent, albeit limited, set of [+hum] reference nouns and adjectives, reminiscent of vestigial gender in the French-based creoles (Valdman 1978; DeGra& 2001). !is feature is also present in its o&shoots, Macau CP and Tugu CP, and it is detected in other Iberian-based creoles. Although unproductive in MCP, this morphology does have a type of functional vitality, as speakers have an awareness of the a#xing. For descriptive purposes, rather than treat natural gender as being lexicalized in these nouns and adjectives, it was hypothesized that the etymological morphemes could be treated as a type of derivation.

Nevertheless, NGM in MCP appears to display traces of a contextual dimension, with a limited variable and categorical matching of NGM items at intra-phrasal and inter-phrasal levels. Initially, it was suggested that this matching could be treated as a purely lexical phenomenon, governed by semantics, and perhaps treated as compounding. However, subsequently, it was demonstrated that the NGM ADJ can be separated from the head N, either by an intensifying ADV or by syntactic move-ment. Also, it was seen that the matching of items with common natural gender can apply between subject NP and a predicative ADJ or N. While these cases could be handled by pragmatic-semantic selection of fully lexicalized items, there is still a minor residue of a syntactic matching visible, at least within the NP.

It is curious that natural gender morphology should be conserved in a consis-tent set of items in MCP, and in certain other Romance-based creoles. Relatively few FLA and SLA studies have identi"ed a role for natural gender in the acquisi-tion of gender morphology. Yet, it may have a role in naturalistic acquisition in contact situations involving particular language interfaces. In this respect, it was noted that distant Indian substrates may have made a contribution to MCP. On the other hand, the presence of NGM in MCP in preference to number morphology is accounted for on grammar-internal bases, and by saliency factors. Gender at-tribution is intrinsic to the N, whereas number attribution is assigned in the head of NumP and its appearance there is a result of phrase construction. Indeed, the FLA and SLA research literature on Spanish and Portuguese provides evidence of the development of morphological gender earlier than morphological number.

Adopting the perspectives of the Interlanguage Hypothesis and Processability !eory, it would seem that we can detect four stages achieved in the acquisition of items bearing gender morphology in SLA input to MCP. Firstly, the forms of nouns and adjectives and determiners in MCP suggest that, in the majority of cases, grammatical gender morphology, just as grammatical number morphology, simply was not acquired and the default form (ex-masculine gender) was adopted in a blanket manner. !is would correspond to a broad word-level early-stage


processing in SLA, devoid of procedures to handle in$ection, and strongly direct-ed by data characteristics in available Portuguese models (semantic and pragmatic saliency, guided by frequency) and e&ected by L1 transfer. Secondly, as category procedures developed and lexical morphology emerged, natural gender morphol-ogy was incorporated, directed by both semantic-pragmatic factors, frequency, and by salient formal aspects of gender in$ection such as in$ected [+hum] heads and double-marked items (i.e. metaphonic gender). Subsequently, at a third stage, items with natural gender morphology began to be involved in intra-phrasal rela-tions in the form of N+ADJ matching, corresponding to a developmental stage in syntax that facilitated intra-phrasal topicalization. Finally, at a fourth stage, it seems that items with natural gender morphology may have achieved a minimal inter-phrasal dimension, showing information exchange between items in the subject and the predicate.

!e diachronic status of the in$ectional morphology in MCP is di#cult to assess without linguistic evidence from earlier centuries. SLA input may have di-rected the development of inherent gender morphology in MCP both in the initial stages and through ongoing contact with Portuguese and Indo-Portuguese creoles.

Abbreviations

? — of uncertain grammaticality f — feminine gender* — ungrammatical hum — humanadj — adjective m — masculine genderadv — adverb MCP — Malacca Creole Portugueseanim — animate n — nounarch — archaic item NGM — natural gender morphologycp — Portuguese lexically-based creole poss — genitivedem — demonstrative um — unmarked for genderdet — determiner

References

Alarcón, Irma. 2004. !e sequential acquisition of L2 Spanish gender marking: Assignment and agreement. Indiana University Linguistics Club Working Papers 4 (7). 1–23.

Andersen, Roger W. 1980. Creolization as the acquisition of a second language as a "rst lan-guage. In Albert Valdmann & A. High"eld (eds.), "eoretical orientations in creole studies, 273–295. San Diego: Academic Press.

Andersen, Roger W. 1984. What’s gender good for, anyway? In Roger W. Andersen (ed.), Second languages: A cross-linguistic perspective, 77–100. Rowley, MA: Newbury House.

Arono&, Mark & Fudeman, Kirsten. 2005. What is morphology? Oxford: Blackwell Publishing.

!"#$%&''()

146 Alan N. Baxter

Bakker, Peter. 2003. Pidgin in$ectional morphology and its implications for creole morphology. In Ingo Plag (ed.), Yearbook of Morphology 2002, 3–33. Dordrecht: Kluwer.

Baptista, Marlyse. 2002. "e syntax of Cape Verdean Creole. Amsterdam/Philadelphia: John Benjamins.

Baxter, Alan N. 1988. A grammar of Kristang (Malacca Creole Portuguese). Canberra: Paci"c Linguistics.

Baxter, Alan N. 1996. Portuguese in the Paci"c and Paci"c Rim. In Stephen A. Wurm, Peter Mühlhäusler & Darrell T. Tryon (eds.), Language Atlas for Intercultural Communication in the Paci$c Rim Area, 299–338. Berlin: Mouton de Gruyter.

Baxter, Alan N. 2005. Kristang (Malacca Creole Portuguese) — a seriously endangered lan-guage. Estudios de Sociolingüística 6 (1). 1–37.

Baxter, Alan N. 2009. O português em Macau: contato e assimilação. In Ana Maria Carvalho (ed.), Português em Contato, 277–312. Madrid/Frankfurt: Iberoamericana/Editorial Vervuert.

Baxter, Alan N. & de Silva, Patrick. 2004. A dictionary of Kristang (Malacca Creole Portuguese) — English. Canberra: Paci"c Linguistics.

Becker, Angelika & Veenstra, Tonjes. 2003. !e survival of in$ection in French-related Creoles — !e role of SLA processes. Studies in Second Language Acquisition 25. 283–306.

Bickerton, Derek. 1988. Creole languages and the bioprogram. In Frederick J. Newmeyer (ed.), Linguistics: "e Cambridge Survey — Vol 2, 268–284. Cambridge: Cambridge University Press.

Bickerton, Derek. 1999. How to acquire language without positive evidence: What acquisi-tionists can learn from creoles. In Michel DeGra& (ed.), Language Creation and Language Change, 49–74. Cambridge, MA: MIT Press.

Booij, Geert. 1995. Inherent vs. contextual in$ection and the split morphology hypothesis. In Geert Booij & Jaap van Marle (eds.) Yearbook of Morphology 1994, 1–16. Dordrecht: Kluwer.

Booij, Geert. 2007. "e grammar of words. Oxford: Oxford University Press.Brito, António de Paula (1967 [1887]) Dialectos crioulos portugueses. Apontamentos para

a gramática do crioulo que se fala na ilha de Santiago de Cabo Verde. In Jorge Morais-Barbosa (ed.), Estudos linguísticos crioulos, 329–404. Lisbon: Academia Internacional de Cultura Portuguesa.

Bybee, Joan. 1985. Morphology — a study in the relation between meaning and form. Amsterdam/Philadelphia: John Benjamins.

Cardoso, Hugo C. 2009. "e Indo-Portuguese language of Diu. Utrecht: Landelijke Onderzoekschool Taalwetenschap.

Clements, J. Clancy. 1992. Foreigner talk and the origins of Pidgin Portuguese. Journal of Pidgin and Creole Languages 7. 75–92.

Clements, J. Clancy. 1996. "e genesis of a language. Amsterdam/Philadelphia: John Benja mins.Clements, J. Clancy & Koontz-Garboden, Andrew. 2002. Two Indo-Portuguese creoles in con-

trast. Journal of Pidgin and Creole Languages 17. 191–236.Corbett, Greville G. 1991. Gender. Cambridge: Cambridge University Press.Corbett, Greville G. 2006. Agreement. Cambridge: Cambridge University Press.Crowley, Terry. 2008. Pidgin and creole morphology. In Silvia Kouwenberg & John Victor Singler

(ed.), "e handbook of pidgin and creole studies, 74–97. Chichester: Wiley-Blackwell.Dalgado, Sebastião R. 1906. Dialecto Indo-Português do Norte. Revista Lusitana 9. 142–166,

193–228.Dalgado, Sebastião R. 1917. Dialecto Indo-Português de Negapatão. Revista Lusitana 20. 40–53.


Dalgado, Sebastião R. 1998. Dialecto Indo-Português de Ceilão. Lisbon: Commissão Nacional para as Comemorações dos Descobrimentos Portugueses.

DeGra&, Michel. 2001. Morphology in creole genesis. In Michael Kenstowicz (ed.), Ken Hale: A life in language, 55–121. Cambridge, MA: MIT Press.

Domínguez, Alberto, Cuetos, Fernando & Seguí, Juan. 1999. !e processing of grammatical gender and number in Spanish. Journal of Psycholinguistic Research 28. 485–498.

Emmerich, Charlotte. 1984. A língua de contato no Alto Xingu — Origem, forma e função. Rio de Janeiro: Universidade Federal do Rio de Janeiro doctoral dissertation.

Emmerich, Charlotte. 1993. O português de contato no parque indígena do Xingu, Mato Grosso, Brasil Central. Estudos Lingüísticos e Literários (UFBa) 13. 57–90.

Fernández-García, Marisol. 1999. Patterns of gender agreement in the speech of second lan-guage learners. In Javier Gutiérrez-Rexach & Fernando Martínez-Gil (ed.), Advances in Hispanic linguistics: Papers from the 2nd Hispanic Linguistics Symposium, 3–15. Somerville, MA: Cascadilla Press.

Fernando, Radin. 2004. Metamorphosis of the Luso-Asian Diaspora in the Malay Archipelago, 1640–1795. In Peter Borschberg (ed.), Iberians in the Singapore-Melaka Area and Adjacent Regions (16th to 18th Century), 161–184. Wiesbaden: Harrassowitz Verlag.

Ferraz, Luís I. 1988. Portuguese creoles of West Africa and Asia. In Glenn G. Gilbert (ed.), Pidgin and Creole Languages — Essays in Memory of John E. Reinecke, 337–360. Honolulu: University of Hawaii Press.

Finnemann, Michael D. 1992. Learning agreement in the noun phrase: !e strategies of three "rst-year students. International Review of Applied Linguistics 30. 121–135.

Forman, Michael D. 1972. Zamboagueño texts with grammatical analysis, a study of Philippine Creole Spanish. Ann Arbor: University Micro"lms International.

Ferreira, José dos Santos. 1996. Papiaçám di Macau. Macau: Fundação Oriente.Franceschina, Florencia. 2005. Fossilized Second Language Grammars. Amsterdam/Philadelphia:

John Benjamins.Gair, James W. & Lust, Barbara C. 1998. Studies in South Asian linguistics: Sinhala and other

South Asian languages. New York: Oxford University Press.Garling, Mrs. 1830. Letter from Mrs. Garling, dated March 28 1830. Malacca — Incoming let-

ters, Box 3. Correspondence of the London Missionary Society. Council for World Mission Archives (ex-London Missionary Society), School of African and Oriental Studies, London.

Godinho, Ana Paula Marques Cleto de Oliveira. 2005. A Aquisição da concordância de plural no sintagma nominal por aprendentes chineses de português língua estrangeira. Lisbon: Universidade de Lisboa doctoral dissertation.

Hancock, Ian F. 1969. !e Malacca Creoles and their language. Afrasian 3. 38–45.Hancock, Ian F. 1973. Malacca Creole Portuguese: a brief transformational account. Te Reo 16.

23–44.Hancock, Ian F. 1975. Malacca Creole Portuguese: Asian, African or European? Anthropological

Linguistics 17. 211–236.Holm, John. 1989. Pidgins and Creoles. Vol. II. Cambridge: Cambridge University Press.Holm, John A. 2000. An introduction to pidgins and creoles. Cambridge textbooks in linguistics.

New York: Cambridge University Press.Holm, John. 2005. Creolization and the fate of in$ections. Paper presented at Conference on

Romanicization World-Wide, University of Bremen.Irmen, Lisa & Knoll, Jochen. 1999. On the use of the grammatical gender of anaphoric pronouns

in German. A comparison between Finns and Germans. Sprache & Kognition, 18. 123–135.

!"#$%&''()

148 Alan N. Baxter

Kester, Ellen-Petra. 1996. On the Nature of Adjectival In5ection. Utrecht: OTS.Kihm, Alain. 2003. In$ectional categories in creole languages. In Ingo Plag (ed.), Phonology and

morphology of creole languages, 333–363. Tübingen: Niemeyer.Knowlton, J. Edgar, Jr. 1964. Malaysian Portuguese. "e Linguist 26. 211–213, 239–241.Krishnamurti, Bhadruraju. 2003. Dravidian Languages. New York: Cambridge University Press.Lucchesi, Dante. 2000. A variação na concordância de género em uma comunidade de fala afro-

brasileira: Novos elementos sobre a formação do português popular do Brasil. Rio de Janeiro: Universidade Federal do Rio de Janeiro doctoral dissertation.

MacGregor, Ian A. 1955. Notes on the Portuguese in Malaya. Journal of the Malayan Branch of the Royal Asiatic Society 67. 57–84.

Masica, Colin P. 1991. "e Indo-Aryan Languages. Cambridge: Cambridge University Press.Matos, Maria Isabel Silva. 2008. A concordância em género nos constituintes do sintagma nomi-

nal: um estudo com aprendentes chineses de Português L2. Macau: Universidade de Macau masters dissertation.

Maxwell, W. George. 1911. Barretto de Resende’s account of Malacca (<1638). Journal of the Straits Branch of the Royal Asiatic Society 60. 1–24.

McWhorter, John. 2005. De$ning Creole. Oxford: Oxford University Press.Mello, Tavares de. 1998. Folclores Ceilonenses. Colectânea de textos do crioulo português de Ceilão.

Lisbon: Commissão Nacional para as Comemorações dos Descobrimentos Portugueses.Montrul, Silvina. 2004. "e Acquisition of Spanish — Morphosyntactic development in monolin-

gual and bilingual L1 acquisition and adult L2 acquisition. Amsterdam/Philadelphia: John Benjamins.

Munteanu, Dan. 1996. El Papiamentu, lengua criolla hispánica. Madrid: Gredos.Pienemann, Manfred. 2000. Psycholinguistic mechanisms in the development of English as

a second language. In Ingo Plag and Klaus P. Schneider (eds.), Language Use, Language Acquisition and Language History: (Mostly) Empirical Studies in Honour of Rüdiger Zimmermann, 99–118. Trier: Wissenscha%licher Verlag Trier.

Pienemann, Manfred (ed.). 2005. Cross-linguistic aspects of Processability "eory. Amsterdam/Philadelphia: John Benjamins.

Plag, Ingo. 2001. !e nature of derivational morphology in creoles and non-creoles. Journal of Pidgin and Creole Languages 16. 153–160.

Plag, Ingo. 2005. Morphology in pidgins and creoles. In Keith Brown (ed.), Encyclopedia of Language and Linguistics, 2nd ed., vol. 8, 304–308. Oxford: Elsevier.

Plag, Ingo. 2008. Creoles as interlanguages: In$ectional morphology. Journal of Pidgin and Creole Languages 23. 114–135.

Rêgo, António da Silva. 1998 [1942]. Dialecto Português de Malaca — e outros escritos. Lisbon: Commissão Nacional para as Comemorações dos Descobrimentos Portugueses.

Roberts, Sarah J. & Bresnan, Joan. 2008. Retained morphology in pidgins: A typological study. Linguistic Typology 12. 269–302.

Schuchardt, Hugo. 1882. Kreolische Studien VI. Über das Indoportugiesische von Cochim. Sitzungsberichte der Kaiserlichen Akademie der Wissenscha#en zu Wien 102. 799–816.

Schuchardt, Hugo. 1883. Kreolische Studien VI. Über das Indoportugiesische von Mangalore. Sitzungsberichte der Kaiserlichen Akademie der Wissenscha#en zu Wien 105. 881–904.

Schuchardt, Hugo. 1889. Beiträge zur Kenntnis dês kreolischen Romanisch VI. Zum Indoportugiesischen von Mahé und Cannanore. Zeitschri# Für Romanische Philologie 13. 516–524.


Schuchardt, Hugo. 1891. Kreolische Studien IX. Über das Malaioportugiesische von Batavia und Tugu. Sitzungsberichte der Kaiserlichen Akademie der Wissenscha#en zu Wien 122. 1–258.

Schwegler, Armin & Green, Kate. 2007. Palenquero (Creole Spanish). In John Holm & Peter L. Patrick (eds.), Comparative Creole Syntax, 273–306. London: Battlebridge Publications.

!omaz, Luis F. 1985. !e Indian merchant Communities in Malacca under the Portuguese rule. In Teotónio R. de Souza (ed.), Indo-Portuguese History — Old Issues, New Questions, 56–72. New Delhi: Concept.

Tomás, Isabel Gonçalves. 2004. O Kristang de Malaca — processos linguísticos e contextos sociais na obsolescência das línguas. Lisbon: Universidade Nova de Lisboa doctoral dissertation.

Valdman, Albert. 1978. Le Créole. Structure, Statut et origine. Paris: Klincksieck.Villiers, John. 1987. Portuguese Malacca and Spanish Manila: two concepts of empire. In

Roderich Ptak (ed.), Portuguese Asia: Aspects in history and economic history (sixteenth and seventeenth centuries), 37–58. Stuttgart: Steiner Verlag.

White, Lydia, Valenzuela, Elena, Kozlowska-MacGregor, Martyna & Leung, Yan-Kit Ingrid. 2004. Gender and number agreement in nonnative Spanish. Applied Psycholinguistics 25. 105–133.

!"#$%&''()

Bazaar Malay topics*

Bao Zhiming and Khin Khin AyeNational University of Singapore / Swinburne University of Technology, Kuching, Malaysia

Bazaar Malay is a Malay-lexi"ed pidgin with a Chinese substratum spoken in the marketplace of Singapore (and elsewhere in Southeast Asia). Although it is no longer a lingua franca in Singapore today, it is nevertheless still spoken by older Singaporeans. Like Chinese and Malay, Bazaar Malay is a topic-prominent language. We document three types of the Bazaar Malay topic construction and show that they are identical to the topic structures found in Chinese. !e degree of convergence in the topic construction between Chinese and Bazaar Malay, and between Chinese and Singapore English, supports the systemic view of substratum transfer.

1. Introduction

Bazaar Malay is a Malay-lexi"ed pidgin which was widely used as a lingua franca in the Malay peninsula and the Indonesian archipelago before the British East Indian Company annexed Singapore as a trading port in 1819. In the early decades of the crown colony, Bazaar Malay continued to play that role, and the British govern-ment o#cers were o%en advised to learn it when they "rst arrived to assume o#ce (Dennys 1878). Unlike Baba Malay, which is the mother tongue of the Peranakan community in the Straits Settlements,1 Bazaar Malay has never had native speak-ers. !e Chinese, Malay, and Indian communities that comprise the population

* We wish to express our gratitude to Umberto Ansaldo, Lionel Wee, Donald Winford, and the anonymous referees for their insightful comments, discussions, and encouragement. Khin Khin Aye also wishes to thank Umberto Ansaldo and Anne Pakir for their guidance in her work on Bazaar Malay. All errors of facts and interpretation are our own.

1. Peranakan is a Malay word that means ‘locally born’. !e Peranakan Chinese, or Babas, are descendents of early settlers from the southern Chinese coast who married local Malay-speaking women (Vaughn 1879, Turnbull 1977, Tan 1988, Rudolph 1998, Ansaldo, Lim, & Mufwene 2007). !e Babas speak a Malay-lexi"ed patois called Baba Malay, which, according to Shellabear (1912, 1913), mixes Malay words with Chinese idiom or grammar. In Singapore, the

!"#$%&''()

152 Bao Zhiming and Khin Khin Aye

of Singapore (and Malaysia) retain their respective mother tongues. From the be-ginning of the 20th century English started to replace Bazaar Malay as the lingua franca of choice in Singapore, and the process accelerated a%er Singapore gained independence in 1965. Although Bazaar Malay is no longer a lingua franca in Singapore, many older Singaporeans can still speak it $uently.

As expected, most of the words in Bazaar Malay come from Malay, without the usual a#xation active in Malay morphology. !ere are words of Chinese ori-gin, such as lu ‘you,’ wa (or gwa) ‘I,’ tauge ‘bean sprouts’, and lang ‘person,’ some of which are also used in Baba Malay and other local languages. Many of the Chinese loanwords are of Hokkien pronunciation, betraying the in$uence of the dialect in the early days of Singapore, and indeed of the region. !ere are also English-derived words in Bazaar Malay (persen ‘percent,’ tayar ‘tyre,’ aksiden ‘accident’). Like the Chinese loanwords, the English loanwords are also in active use in the vocabularies of the languages spoken in Singapore.

Broadly speaking, the basic word order of Bazaar Malay is SVO, consistent with other pidgins, and with the contributing languages — Malay, the lexi"er, and Chinese, the main substratum language. !e basic clause structure is exempli"ed by the data shown in (1).2

(1) a. Saya sudah tujoh puluh lima umur 1sg perf seven ten "ve age ‘I am already seventy "ve years old.’ b. Saya tengok itu orang jantan. I see this person male ‘I saw the man.’ c. Baru tahu badan semua sakit sampai tak boleh tahan. recently know body all sick until not can bear ‘Recently (he) noticed that the whole body ached so much he could not

bear (it).’

community is among the "rst to shi% to English from Baba Malay. !e website peranakan.org.sg contains interesting information about the community.

2. !e Bazaar Malay data reported in this paper were collected by Khin Khin Aye between 2002 and 2004 from 10 informants, aged between 49 and 77 at the time of data collection. Four of the informants are Chinese, four Indian, and two Malay. !e data gathering sessions were conduct-ed in community centers and co&ee shops in Geylang, a Malay enclave during the colonial days. Details of the data collection are explained in Aye (2005). !e data are cited without regard to the ethnicity of the source, and are understood by all our informants. We assume that the topic construction is a stable grammatical construction of Bazaar Malay, and is therefore part of the structural repertoire of a $uent speaker.

Bazaar Malay topics 153

d. Dia angkat ini besi. 3sg li% this iron ‘He li%ed this iron.’

As far as the basic SVO word order is concerned, there is not much di&erence among the languages in contact.

Like Baba Malay, Bazaar Malay has strong Chinese in$uence in its grammar, which has not escaped the notice of keen observers of the vernacular (Dennys 1878, McHugh 1948, Winstedt 1957, Asmah 1982, 1987, Collins 1987, Lim 1988, Adelaar & Prentice 1996, Aye 2005, Rekha 2007). Winstedt (1957) lists many Chinese-derived features, two of which are displayed in (2) (mod, modi"er).3

(2) a. Dulu punya pasar old mod market ‘old market’ b. Saya kasi dia tahu awak datang I give 3sg know you come ‘I let him know that you have come.’

In (2a), the Malay verb punya ‘own, possess’ is used as a grammatical morpheme that links the pre-modi"er dulu with the nominal head pasar, modeled on the Hokkien particle e (Mandarin de). Here, it has lost its lexical meaning, and its function in Bazaar Malay is to introduce modi"ers. In (2b), kasi ‘give’ is used in the causative sense, consistent with the usage pattern of the Hokkien verb ho (Mandarin gei) with the same lexical meaning. Not surprisingly, these features also appear in Baba Malay; see Shellabear (1912, 1913), Pakir (1986), Lim (1988), Ansaldo & Matthews (1999), and Lee (1999).4

3. !e in$uence, of course, extends to pronunciation as well. In Bazaar Malay, words such as tahu ‘know’ and mahu ‘want’ are pronounced without the intervocalic h, and banyak ‘many’ is pronounced as manyak. Despite the obvious di&erences in pronunciation, we adopt standard Malay spelling conventions in spelling Malay-derived words from our database of Bazaar Malay.

4. !e Chinese in$uence in the use of punya to express the possessive has been noted by ear-ly students of Baba Malay and Malay. Shellabear (1912: 6–7) warns: ‘!e frequent use of pu-nya, however, is a Chinese idiom, and, though common with the Straits-born Chinese, should be avoided by Europeans who wish to speak with any approach to correctness.’ Straits-born Chinese are the Peranakans.

!"#$%&''()


2. Topic prominence

In this paper, we focus on topic prominence, another salient feature of Bazaar Malay. Although the topic construction is common among pidgins and creoles generally (Holm 2000, Winford 2003), Bazaar Malay topic prominence is heavily in$uenced by Chinese, the main substratum language.5 We document the whole range of topic structures in our Bazaar Malay database, and show that the Bazaar Malay topic structures are point-to-point identical with those of Chinese. Our analysis provides further empirical evidence in support of the role of grammatical system in substratum transfer (see Bao 2005).

!ree structural properties have been associated with Chinese topic promi-nence in the literature (Chao 1968, Chafe 1976, Li & !ompson 1976, 1981, Xu & Langendoen 1985, Bao 2001, Bao & Lye 2005). !ese are exempli"ed in (3):

(3) a. English-style Nà-ge nü"-hái, Zhāngsān xĭhuān tā/e that girl Zhangsan like her/e ‘!at girl, Zhangsan likes her/e.’ b. Chinese-style Sh#igu$, Zhāngsān xĭhuān pínggu$ fruit Zhangsan like apple ‘As for fruits, Zhangsan likes apples.’ c. Bare conditional Nĭ xĭhuān tā, nĭ yīnggāi zìjĭ gàoshù tā you like her you should self tell her ‘If you like her, you should tell her yourself.’

!e topics are set in italics. !e data are transcribed in pinyin in accordance with Mandarin pronunciation.6 In the English-style topic structure, the topic is

5. As pointed out by two anonymous referees, typologically Malay is topic-prominent as well, which weakens the argument that the topic prominence of Bazaar Malay derives from Chinese. We are not in a position to discuss Malay topic structures here. Our focus is on the structural manifestation of topic prominence in Chinese, shown in (3). Any degree of convergence in the topic construction between Chinese and Malay will only facilitate the stabilization of the topic structures in Bazaar Malay.

6. !e numerically dominant native dialects of early Chinese immigrants to Singapore are Cantonese and the Southern Min dialects of Hokkien and Teochew. !e role of Mandarin in early language contact in Singapore should not be dismissed entirely, since it was the medium of in-struction in Chinese vernacular schools before the government introduced nation-wide English-medium education. As far as the topic construction is concerned, there is no di&erence among the dialects.


semantically related to a resumptive or null pronominal in the comment, marked respectively by tā ‘her’ or e in (3a). In generative linguistics, the English-style topic structure is derived through movement, leaving behind a trace marked by e. !e process is o%en referred to as topicalization. In the Chinese-style topic structure, the comment contains no pronominal, null or otherwise, to which the topic is related, as shown in (3b). For this reason, the Chinese-style topic structure resists movement analysis (Xu & Langendoen 1985). (3c) is an example of what Bao & Lye (2005) call bare conditional, where the conditional clause is not marked by an if-word.

Chinese conditionals are typically marked by a host of adverb-like expressions such as yàoshì, rúgu$, tángruò, or jiărú, all of which can be glossed with English ‘if ’; see Chao (1968) and Li & !ompson (1981). !e position of these if-words is not "xed — they may appear clause-initially (4a), or clause-medially (4b), or be missing (4c) (examples cited from Bao & Lye 2005).

(4) a. Yàoshì kăoshì yánqī, w$ jiù bù néng qù Yīngguó if exam postpone I then not able go England ‘If the exam postpones, then I won’t be able to go to England.’ b. Kăoshì yàoshì yánqī, w$ jiù bù néng qù Yīngguó exam if postpone I then not able go England ‘If the exam postpones, then I won’t be able to go to England.’ c. Kăoshì yánqī, w$ jiù bù néng qù Yīngguó exam postpone I then not able go England ‘If the exam postpones, then I won’t be able to go to England.’

Such $exibility in the use of if-words in the Chinese conditional is not found in English, where the conditional clause is introduced by if in clause-initial position, and the bare conditional construction is used only for special e&ect (Say that again and you’re $red; see Huddleston & Pullum 2002). Chao (1968) uses topic and the associated notion comment as pragmatic notions, and treats all Chinese condi-tionals, marked or bare, as topics. For descriptive convenience we will discuss bare conditionals only. Following Xu & Langendoen (1985), we treat topic and com-ment as structural constructs. So the crucial di&erence between English-style and Chinese-style topic structures can be expressed as follows (top, topic; s, comment):

(5) a. English-style: […]TOP […e…]S b. Chinese-style: […]TOP […]S

top is the syntactic position that can be occupied by all major phrasal categories. !e bare conditional structure obtains when the topic is interpreted as a condi-tional, without the use of any of the if-words. Structurally, it is identical to the Chinese-style topic, since the conditional clause typically does not originate with-in the comment clause, now interpreted as the consequent.

!"#$%&''()


!e three types of topic structure shown in (3) are diagnostic of Chinese topic prominence. We now proceed to show that Bazaar Malay has precisely these three types.

3. English-style topic structure

We "rst look at the English-style topic structure, which is exempli"ed below:

(6) a. [Anjing]TOP [dia gigit sama dia punya tangan]S dog 3sg bite with 3sg mod hand ‘!e dog, it bit his hand.’ b. [Itu gemuk punya orang]TOP [dia cakap dia boleh angkat naik]S that fat mod person 3sg speak 3sg can li% up ‘!e fat man, he said he could li% (it) up.’ c. [Kerja punya pasal]TOP [dia tanya e]S work mod matter 3sg ask ‘(A) work-related matter, he asked (about).’ d. [Itu]TOP [dia tak suka e]S that 3sg not like ‘She does not like it. e. [Itu panggil apa]TOP [saya pun tak tahu e]S that call what I also not know ‘What it is called, I also don’t know.’ f. [Apa sayur dia cakap]TOP [saya tahu e]S what vegetable 3sg speak I know ‘What vegetables he mentions, I know.’ g. [Di cina]TOP [sudah ada bini e]S in China already have wife ‘In China, (he) had already had a wife.’

Punya is the grammatical particle that introduces pre-modi"cation, see (2a). In (6a,b), the comment contains the resumptive pronoun dia; in (6c–g), it contains the null pronominal e. In terms of grammatical category, top is occupied by a noun phrase in (6a–d), by a sentence in (6e,f) and by a preposition phrase in (6g). In terms of grammatical function, top is the subject in (6a,b), the object in (6c–f) and the adjunct in (6g).7 !e English-style topics have their direct counterparts in English, as indicated by the English translations in (6).

7. We assume that the basic constituent order of a Malay verb phrase is as follows:

V – (Object) – Locative – Temporal


4. Chinese-style topic structure

Unlike English-style topic structure, the topic in Chinese-style topic structure does not originate in the comment. We have seen one example in (3b). More ex-amples are given in (7).

(7) a. Nèi-xiē shùmù shùshēn dà. those tree tree-trunk big ‘!ose trees, the trunks are big.’ b. Rénjiā shì fēng nián others be bumper year ‘As for those people, (it) is a bumper year.’

(7a) is cited from Li & !ompson (1981: 462) and (7b) from Chao (1968: 71). (Tone marks are ours). In (7a), the subject of the comment sentence is part of the topic; in (7b), the topic provides the reference frame for the interpretation of the comment. !e two semantic relationships between the topic and comment exem-pli"ed in (7) are the typical relationships that characterize the Chinese-style topic structure.

Chinese-style topic structure is also attested in Bazaar Malay (part, particle):

(8) a. [Bangsa ni Singapore]TOP [Teochew, Hokkien, Fouchow, Kwanton race in Singapore Teochew Hokkien Foozhou Cantonese

sini datang bangsa punya orang lah]S here come race mod person part ‘As for races of Singapore, Teochew, Hokkien, Foozhou, Cantonese are

immigrants.’ b. [Singapore punya orang]TOP [terlinga dia dengar mana yang barang Singapore mod people ear 3sg hear where that goods

murah, dia pergi]S cheap 3sg go ‘As for Singaporeans, if his ear hears where things are cheap, he will go.’

!e locative and temporal adjuncts appear a%er the main verb, and the object, if any. !e ex-ample below illustrates this order (Mintz 1994: 37):

So"ah ada pergi ke mana-mana besok? So"ah have go to anywhere tomorrow ‘Is So"a going anywhere tomorrow?’

For this reason, we analyze the prepositional phrase di Cina ‘in China’ to be topicalized from its canonical position marked by e.

!"#$%&''()


c. [Itu barang ah]TOP [ada kali mahal, ada kali murah, itu that goods part have time expensive have time cheap that pasar sana]S market there ‘As for those goods, sometimes expensive, sometimes cheap, that

market.’ d. [Tapi saya minya tempat keraja]TOP [bukan Melayu lang Burma, but I mod place work no Malay person Burma

Indonesia, Bangladesh, China, India]S Indonesia Bangladesh China India ‘But, as for my work places, (there are) no Malay people, only Burmese,

Indonesians, Bangladeshis, Chinese, and Indians.’

In (8d), minya is a variant of punya, and lang ‘person’ is a Hokkien loanword. As the English translations suggest, none of the topics in (8) can be plausibly analyzed as originating in the respective comment clause. Instead, the comment provides information about the topic. In (8b) the comment clause, which may be given the structure shown below, contains a topic of its own:

(9) [Terlinga dia dengar mana yang barang murah]TOP [dia pergi]S ear 3sg hear where that goods cheap 3sg go ‘(If) his ear hears where things are cheap, he will go.’

!e topic here is interpreted as a conditional clause, even though it is not so marked (see the bare conditional construction below). In (8c), the subject of the comment clause, itu pasar sana ‘that market there,’ appears at the end of the clause.

5. Bare conditionals

Conditionals in Bazaar Malay are typically introduced with the Malay word kalau. Two examples are shown below:

(10) a. Kalau itu Hokkien, kita cakap Hokkien sama dia lah if that Hokkien we speak Hokkien with 3sg part ‘If that (person) is Hokkien, we speak Hokkien with him.’ b. Kalau mahu minum, minum kope satu paket if want drink drink co&ee one packet ‘If (we) want to drink, (we) drink one cup of co&ee.’

However, the conditional marker kalau need not be used to introduce the condi-tional clause. Indeed, the bare conditional construction is very productive in our database. We have seen one example in (9); more data follow.


(11) a. [Kita cerita perkara sedikit]TOP boleh we story matter little can ‘(If) we talk little, (we) can.’ b. [Dia tahu lu bukan Hokku lang]TOP tak mahu cakap ah, tak 3sg know you no Foozhou person not want speak part not

mahu cakap lain punya bangsa want speak other mod race ‘(If) he knows you are not a Foozhou, (he) does not want to speak, does

not want to speak to other races.’ c. [Duluk sana leh]TOP apa-apa pun boleh dapat sit there part anything also can get ‘(If you) stay on, (you) can get everything.’ d. [Lu mahu sekarang]TOP lu tengok you want now you look ‘(If) you want now, you look.’ e. [Ada orang mahu masuk]TOP susah lah have person want enter di#cult part ‘(If) people want to come in, (it’s) di#cult.’ f. [Tak tahu ah]TOP tengok orang cakap apa, lu tahu diaorang not know part look person speak what you know they cakap ah speak part ‘(If you) don’t know, check what people say, you know what they say.’

In (11), we analyze the bare conditional as top, and the consequent as comment, which is not bracketed. Naturally, these examples are felicitous if the topic clause is marked by kalau. !e bare conditionals may also be followed by particles that typi-cally mark topics, such as leh in (11c) and ah in (11f). !is is to be expected, since structurally they occupy the top position, allowing the particles to be attached.

6. Multiple topics and multiple comments

In the preceding sections we documented the three basic types of topic construc-tion in Chinese and Bazaar Malay: the English-style topic structure, Chinese-style topic structure, and the bare conditional as topic. !e data we have seen contain single topics and single comments, even though the comment clause may itself be analyzable as topic-comment, as in (9). In topic prominent languages, a sentence may contain more than one topic to be used with a single comment clause, and more than one comment clause commenting on a single topic. Relevant examples are shown below (Li & !ompson 1981, Xu & Langendoen 1985, Bao 2001):

!"#$%&''()


(12) a. Zuótiāni, Lĭ xiānshēngj, w$ ei kànjiàn ej le. yesterday Li Mr. I see asp ‘Yesterday, Mr. Li, I saw.’ b. Zhèjiàn shìi, Lĭ xiānshēngj, w$ gàosù ej ei guo this event Li Mr. I tell asp ‘!is event, Mr. Li, I have told about.’ c. Zhè-kē shù, yè duō, huā shăo this-cl tree leaf many $ower few ‘!is tree, the leaves are many, the $owers are few.’ d. Nèi-kuài tiáni, dàozi zhăngde hĕn dà, suóyĭ ei hĕn zhíqián that-cl land rice grow very big so very valuable ‘!at piece of land, rice grows very big, so (it) is very valuable.’

(12a,b) contain two topics, and (12c,d) two comments. Note that both (12a,b) are English-style topic structure, with the topics originating in the comment clause. (12c) is Chinese-style, whereas (12d) is a mixture of the English-style (the "rst comment clause) and Chinese-style (the second comment clause).

As in Chinese, Bazaar Malay employs clauses with multiple topics and mul-tiple comments as well. !e multiple-topic data are displayed in (13), and the mul-tiple-comment data in (14).

(13) a. [Bazaar Malay]i [itu jam]j, [semua boleh cakap ei di kampong ej]S Bazaar Malay that time all can speak in village ‘Bazaar Malay, at that time, all could speak in the village.’ b. [Pasal minya lu]i, [saudagar pergi datang pergi datang]j matter mod you merchant go come go come [dia tanya sama ei ej]S 3sg ask with ‘About you, from the merchants who went to and fro, he asked.’

(14) a. [Toto]TOP [mana-mana pun ada e]S, [e bukan di satu tempat]S Toto anywhere also have not in one place ‘Toto, (it is) everywhere, not in one place.’ b. [itu perempuan]TOP [e banyak jahat]S, [tak boleh tahan e]S that woman many vicious not can endure ‘!at woman, (she) was vicious, (I) could not stand (her).’ c. [Orang tua]TOP [e pergi angkat payong ah]S, [e pukul lumbu ah]S person old go carry umbrella part hit bu&alo part ‘!e old man, (he) le% carrying an umbrella, hit the bu&alo.’ d. [Niaga]TOP [Indonesia punya barang ambil keluar Singapore]S, business Indonesia mod goods take out Singapore


[Singapore punya barang ambil masuk Indonesia]S Singapore mod goods take enter Indonesia ‘As for business, Indonesian goods were exported to Singapore,

Singapore goods were imported into Indonesia’.

In (13) and (14), multiple topics are di&erentiated with subscripts, and the null pronominals (ei and ej) appear in the canonical positions of the topics, see foot-note 7. In some cases, the same sentence may contain more than one topic and more than one comment as in the following example:

(15) [Dia cakap apa]TOP [Singapore panggil apa]TOP [dia tahu e]S, 3sg speak what Singapore call what 3sg know [saya tahu e lah]S I know part ‘What he says, what Singapore calls, she knows, I know.’

In (15), the two topics serve as the object within each of the two comment clauses.

7. Comparing topic prominence in Singapore English

In the preceding sections, we have demonstrated the close "t in topic prominence between Bazaar Malay and Chinese — the whole range of topic structures in Chinese is transferred from Chinese to Bazaar Malay. !e same topic structures are found in Singapore English, the vernacular variety that is now being acquired as mother tongue by a growing segment of Singapore’s population (cf. Gupta 1994). It has long been recognized that Singapore English is topic prominent (cf. Platt & Weber 1980). !e relevant data are displayed below.

(16) a. English-style topic structure Certain medicine, we don’t stock in our dispensary. (Platt & Weber

1980: 73) b. Chinese-style topic structure My family everybody is educated in English. ‘In my family everybody is educated in English.’ (Platt, Weber, & Ho

1983: 47) c. Bare conditionals Don’t care lah. Want to eat, eat; don’t want to eat, then don’t eat. ‘Don’t worry. If you want to eat, eat; if you don’t want to eat, then don’t

eat. (Bao & Lye 2005: 283) d. Multiple-topic structure One time, the 5ats, nobody want. (Platt, Weber & Ho 1983: 48)

!"#$%&''()


e. Multiple-comment structure "at $sh, I bought last week, spoiled already. (Bao 2001: 307)

For detailed description and analysis, see the references cited above.!e clustering e&ect exhibited in the topic prominence of Bazaar Malay and

Singapore English raises interesting questions about the nature of substratum transfer and the role of the contact ecology. Most studies of substratum transfer in the contact linguistics literature focus on identifying the source of individual novel lexical and grammatical features in the contact language; for example, see Muysken & Smith (1986), Mufwene (1990, 1993), and Lefebvre (1998). Unfortunately the clustering tendency in substratum transfer has not received its due attention. !e convergence in topic prominence between Chinese and Bazaar Malay (and Singapore English) supports the argument, proposed by Bao (2005) and Bao & Lye (2005), that the grammatical subsystem is the locus of substratist explanation. Systemic substratum transfer takes place in contact environments where the con-tact language interacts with all the input languages throughout its history. !is is the case with both Bazaar Malay and Singapore English.

!roughout its short history, Singapore has had the same mix of languages, and the same sociolinguistic relations among the languages; see Lim (2007) for a recent view on the history of the language situation in Singapore. !e four ‘of-"cial’ languages, English, Chinese, Malay, and Tamil, have had a constant pres-ence in Singapore, although their communicative functions have varied over time. Bazaar Malay and Singapore English emerged and developed in the same linguis-tic ecology, albeit at di&erent times. For both Bazaar Malay and Singapore English, Chinese constitutes the main linguistic substratum. !e early Chinese dialects are Southern Min (Hokkien, Teochew, Hainanese), Hakka, and Cantonese, and since the turn of the 20th century, Mandarin. All the contributing languages — the substratum languages and the lexi"ers — are spoken along with the developing contact languages, albeit for di&erent communicative purposes. Bazaar Malay and Singapore English are constantly being shaped by the same linguistic resources in the contact ecology.8 !is crucial contact condition ensures robust substratum

8. !e systemic transfer of Chinese topic prominence to Bazaar Malay and Singapore English is complete, as we have seen. In other cases of grammatical transfer, the lexi"er language plays an important role in shaping the eventual outcome. Bao (2005) shows that the aspectual system of Singapore English is not identical to the Chinese system or to the English system, the two main input languages. It is in fact shaped by both languages: the Singapore English aspectual system is the Chinese system "ltered through English morphosyntax. Lee (2009) explores the grammar and use of ada ‘have’ in Baba Malay, got in Singapore English, and u ‘have’ in Chinese (Hokkien), and shows that although ada and got have the same substratum source in u, their usage patterns are di&erent, and the di&erence can be explained in terms of the "ltering e&ect of the respective lexi"er languages.


in$uence, resulting in the transfer of the entire cluster of topic structures from Chinese to Bazaar Malay and Singapore English.

Ansaldo (2004: 132) argues for a more nuanced linguistic matrix for Singapore English, shown below:

(17) a. Lexi"er: English vernacular b. Substrate: restructured Malay (Bazaar or Baba Malay) c. Early adstrates: Min varieties: Hokkien, Teochew, Hainanese d. Late adstrates: Yue varieties: Cantonese

To this list we may add Malay, Tamil, Hindi, and other Indian languages, as well as minor Chinese dialects such as Hakka and Hokchia, as early adstrates. In this matrix, Bazaar Malay and Baba Malay are placed as the linguistic substratum to Singapore English, with the Chinese dialects (and other languages) forming the adstratum. It is not clear why Cantonese is classi"ed as a late adstrate, given the fact that Cantonese-speaking immigrants arrived at roughly the same time as oth-er immigrants, and made up roughly 19% of the Chinese community as early as the late 19th century (Bao 2001: 283).

!ere is no doubt that the languages listed in (17), Malay, Tamil, and Indian languages, have been active in the speech community of Singapore to varying degrees. But the distinction between the substratum and adstratum of Singapore English is not easy to maintain for two reasons. First, English has always been the dominant language, with the rest assuming less important roles in the political and economic life of Singapore. As far as Singapore English is concerned, all these languages form the linguistic substratum. Second, the novel linguistic features in Singapore English are mostly Chinese in origin. Although Bazaar Malay and Baba Malay predate Singapore English, they have made negligible contribution to the grammar of Singapore English. For Chinese grammatical constructions that have transferred to Bazaar Malay, Baba Malay, and Singapore English, we would expect that they are transferred to Singapore English directly from Chinese, rather than indirectly through Bazaar Malay or Baba Malay.

In this connection, let us consider punya, which we "rst saw in (2) (see also footnote 4), and one in Singapore English. It has been shown beyond doubt that the punya construction in Baba Malay and Bazaar Malay is appropriated from Chinese (Hokkien e, Mandarin de) (Shellabear 1913, Pakir 1986, Aye 2005). If Baba Malay or Bazaar Malay were the substratum for Singapore English, one would expect the punya construction to transfer to Singapore English, to be re-alized as the one construction. !rough careful analysis of the constructions in question, we can show that this is not the case. One crucial di&erence between the two constructions is that the one construction is productively used for emphasis in Singapore English (Gupta 1992, Bao 2009), but the punya construction does not

!"#$%&''()


have the emphatic use in Baba or Bazaar Malay. One specimen of the emphatic one is shown below (Gupta 1992: 332):

(18) Senior never come and help one ‘Seniors did not come and help!’

If the punya construction is the substratum source of the one construction, we need to explain how one acquires the grammatical function as an emphatic mark-er. Prima facie, the most reasonable explanation is that Singapore English one is modeled directly on Chinese de, which has the emphatic use, rather than indi-rectly through the punya construction.

Ansaldo’s (2004) substrate and adstrate classi"cation is based on the histo-ries of the languages in the contact ecology of Singapore English, rather than on the linguistic e&ect of the distinction on Singapore English. Two historical facts are important. First, Bazaar Malay was the lingua franca before Singapore English took over that function; second, the Peranakan community is among the "rst to switch to English. !ese two facts could potentially give the restructured Malay varieties an opportunity to exert on the emerging Singapore English in$uence out of proportion to their number or position within the early Singaporean commu-nity, as predicted by the Founder Principle (Mufwene 2001). As we have shown, the founder e&ect in Singapore English is obscured by the continued presence of the very languages that have contributed to Bazaar Malay and Baba Malay.

8. Conclusion

In the preceding pages we have demonstrated the complete convergence in top-ic prominence between Chinese and Bazaar Malay. !e salient properties of the Bazaar Malay topic prominence are summarized in (19).

(19) Topic structures in Bazaar Malay a. English-style topic structure b. Chinese-style topic structure c. Bare conditionals d. Multiple-topic structure e. Multiple-comment structure

!is degree of convergence is also observed in Singapore English, which shares the same linguistic substratum as Bazaar Malay.

It is not surprising that Bazaar Malay, like Singapore English, has such robust Chinese-derived topic structures. Speakers of Bazaar Malay maintain their na-tive languages. Our Chinese informants speak their native dialects and Mandarin


as the primary means of communication at home, and Bazaar Malay as a lingua franca between people who do not share a common language. Moreover, Malay, especially spoken Malay, is also topic prominent, as noted by Ra&erty (1987). !e broad typological proximity between the two contributing languages — Chinese the substratum, and Malay the lexi"er — facilitates the transfer of the topic struc-tures qua system from Chinese to Bazaar Malay, and the subsequent stabilization of the transferred grammatical system in Bazaar Malay. Bao (2005) argues for a systemic approach to substratum transfer, which places the explanatory locus on the grammatical system. !e Bazaar Malay topic prominence provides further evi-dence in support of this approach. !e typological closeness between Chinese and Malay with respect to topic prominence allows the entire array of Chinese topic structures to "lter through and become part of the grammar of Bazaar Malay.

References

Adelaar, K. Alexander & David J. Prentice. 1996. Malay: Its history, role and spread. In Stephen A. Wurm, Peter Mühlhäusler & Darell T. Tryon (eds.), Atlas of languages of intercultural communication in Paci$c, Asia and the Americas, Volume II, 6736–6793. Berlin: Mouton de Gruyter.

Ansaldo, Umberto. 2004. !e evolution of Singapore English: Finding the matrix. In Lisa Lim (ed.), Singapore English: A grammatical description, 127–149. Amsterdam: John Benjamins.

Ansaldo, Umberto & Stephen J. Matthews. 1999. !e Minnan substrate and creolization in Baba Malay. Journal of Chinese Linguistics 27(1).38–68.

Ansaldo, Umberto, Lisa Lim & Salikoko S. Mufwene. 2007. !e sociolinguistic history of the Peranakans. What it tells us about ‘creolization’. In Umberto Ansaldo, Stephen Matthews & Lisa Lim (eds)., Deconstructing creole [Typological Studies in Language], 203–226. Amsterdam: John Benjamins.

Asmah, Haji Omar. 1982. Language and society in Malaysia. Kuala Lumpur: Dewan Bahasa dan Pustaka.

Asmah, Haji Omar. 1987. Malay in its sociocultural context. Kuala Lumpur: Dewan Bahasa dan Pustaka.

Aye, Khin Khin. 2005. Bazaar Malay: History, grammar and contact. National University of Singapore dissertation.

Bao, Zhiming. 2001. !e origins of empty categories in Singapore English. Journal of Pidgin and Creole Languages 16(2).275–319.

Bao, Zhiming. 2005. !e aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics 41(2).237–267.

Bao, Zhiming. 2009. One in Singapore English. Studies in Language 33(2).338–365.Bao, Zhiming & Lye Hui Min. 2005. Systemic transfer, topic prominence, and the bare condi-

tional in Singapore English. Journal of Pidgin and Creole Languages 20(2).269–291.Chafe, Wallace L. 1976. Givenness, contrastiveness, de"niteness, subjects, topics, and points of

view. In Charles N. Li (ed.), Subject and topic, 25–55. New York: Academic Press.Chao, Yuen Ren. 1968. A grammar of spoken Chinese. Berkeley: University of California Press.

!"#$%&''()


Collins, James T. 1987. Malaysian and Bazaar Malay: Polarity, continuity and communication. In Asmah Haji Omar (ed.), National language and communication in multilingual societies, 151–174. Kuala Lumpur: Dewan Bahasa dan Pustaka.

Dennys, Nicholas B. 1878. Handbook of Malay Colloquial, as spoken in Singapore: Being a series of introductory lessons for domestic and business purposes. London: Trubner and Co.

Gupta, Anthea F. 1992. Contact features of Singapore Colloquial English. In Kingsley Bolton & Helen Kwok (eds.), Sociolinguistics today: International perspectives, 323–345. London: Routledge.

Gupta, Anthea F. 1994. "e step-tongue: Children’s English in Singapore. Clevedon: Multilingual Matters.

Holm, John. 2000. An introduction to pidgins and creoles. Cambridge: Cambridge University Press.

Huddleston, Rodney D. & Geo&rey K. Pullum. 2002. "e Cambridge grammar of the English language. New York: Cambridge University Press.

Lee, Gwyneth A. Mae-En. 1999. A descriptive grammar of spoken Peranakan. National University of Singapore honors thesis.

Lee, Nala Huiying. 2009. One substrate, two lexi$ers and the lexi$er e!ect. National University of Singapore MA dissertation.

Lefebvre, Claire. 1998. Creole genesis and the acquisition of grammar: "e case of Haitian creole. Cambridge: Cambridge University Press.

Li, Charles N. & Sandra A. !ompson. 1976. Subject and topic: A new typology of language. In Charles N. Li (ed.), Subject and topic, 457–594. New York: Academic Press.

Li, Charles N. & Sandra A. !ompson. 1981. Mandarin Chinese: A functional reference grammar. Berkeley: University of California Press.

Lim, Lisa. 2007. Merges and acquisitions: on the ages and origins of Singapore English particles. World Englishes 26(4).446–473.

Lim, Sonny. 1988. Baba Malay: !e language of the ‘Straits-born’ Chinese. Papers in Western Austronesian Linguistics No. 3 (Paci"c Linguistics Series A Occasional Papers No. 78): 1–61.

McHugh, James Noel. 1948. A handbook of spoken ‘bazaar’ Malay. Singapore: Kelly & Walsh.Mintz, Malcolm W. 1994. A student’s grammar of Malay & Indonesian. Singapore: EPB Publishers.Mufwene, Salikoko S. 1990. Transfer and the substrate hypothesis in creolistics. Studies in

Second Language Acquisition 12(1).1–23.Mufwene, Salikoko S. (ed.). 1993. Africanisms in Afro-American language varieties. Athens: !e

University of Georgia Press.Mufwene, Salikoko S. 2001. "e ecology of language evolution. Cambridge: Cambridge University

Press.Muysken, Pieter & Norval Smith (eds.). 1986. Substrata versus universals in creole genesis.

Amsterdam: John Benjamins.Pakir, Anne. 1986. A linguistic investigation of Baba Malay. Honolulu: University of Hawaii

dissertation.Platt, John T. & Heidi Weber. 1980. English in Singapore and Malaysia: Status, features, functions.

Oxford: Oxford University Press.Platt, John T., Heidi Weber & Mian Lian Ho. 1983. Varieties of English around the world:

Singapore and Malaysia. Amsterdam: John Benjamins.Ra&erty, Ellen. 1987. Word order in intransitive clauses in High and Low Malay of the late nine-

teenth century. In Russell S. Tomlin (ed.), Coherence and grounding in discourse: Typological studies in language, 361–382. Amsterdam: John Bejamins.


Rekha, Sasi. 2007. A description of Singapore Indian Malay, a pidgin’s pidgin. National University of Singapore MA dissertation.

Rudolph, Jürgen. 1998. Reconstructing identities: A social history of the Babas in Singapore. Aldershot, Hants: Ashgate.

Shellabear, William G. 1912. A practical Malay grammar, 3rd edition. Singapore: !e Methodist Publishing House.

Shellabear, William G. 1913. Baba Malay: An introduction to the language of the Straits-born Chinese. Journal of the Straits Branch of the Royal Asiatic Society 65.49–63.

Tan, Chee Beng. 1988. "e Baba of Melaka: Culture and identity of a Chinese Peranakan com-munity in Malaysia. Petaling Jaya: Pelanduk Publications.

Turnbull, C. Mary. 1977. A history of Singapore 1819–1975. Singapore: Oxford University Press.Vaughan, Jonas D. 1879. "e manners and customs of the Chinese of the Straits Settlements.

Singapore: !e Mission Press. (Re-published in 1971 by Oxford University Press, Kuala Lumpur).

Winford, Donald. 2003. An introduction to contact linguistics. Malden, MA.: Blackwell.Winstedt, Richard O. 1957. Colloquial Malay: A Grammar. Singapore: Marican & Sons.Xu, Liejiong & D. Terence Langendoen. 1985. Topic structures in Chinese. Language 61(1).1–27.

!"#$%&''()

Index

!"#$%&''()

170 Index

Pidgins and Creoles in Asia

Documents

Transcript of Pidgins and Creoles in Asia