Population Movements and Contacts

POPULATION MOVEMENTS AND CONTACTS IN LANGUAGE
EVOLUTION1
Salikoko S. Mufwene
University of Chicago
1. Introduction
Publications such as Thomason & Kaufman (1988) and Thomason (2001, 2002) have perpetuated,
without fundamental support, the distinction between internally- and externally-motivated language
changes. They suggest that the two kinds of changes are different in nature and that those induced
by language contact have not contributed to language speciation in the same way as those owing
(principally) to language-internal mechanisms. Putatively, genetic classifications should be based
only on correspondences suggested by internally-motivated change. Thus, they conclude that
creoles cannot be classed genetically, because, in their case, contact was so extensive that the
comparative method cannot be applied to them in informative ways. This disjoint view of language
diversification, which has treated creoles as children out of wedlock (Mufwene 1997b, 2001), is so
deeply entrenched in linguistics that it is repeated both in general introductions to linguistics and in
historical linguistics textbooks. When the latter kinds of books cover the development of creoles at
all as part of their subject matters, they also claim that these vernaculars are genetically
“exceptional,” if not unnatural, because they have not emerged in the “usual” and/or “natural” way.
See, e.g. Hock & Joseph (1996, critiqued in Mufwene 1998, 2001).2
Heine & Kuteva (2005) are rather exceptional in arguing that Europe is a linguistic area, one in
which language contact accounts for the wide diffusion of several grammatical features, suggesting
that the correspondences that obtain among several Indo-European languages are not necessarily
due to common inheritance. It can be concluded from the book that one must know a substantial
amount of the histories of particular populations before determining whether their shared linguistic
materials are due to diffusion through contact or to shared ancestry. Thus, the comparative method
can in principle not be used unilaterally, without taking into account the histories of the relevant
populations, in order to determine whether or not two language varieties are genetically related (see,
1
I am grateful to Peter Bakker for helpful comments on an earlier version of this paper, especially for
recommending it for publication despite our fundamental differences on whether or not creoles are
genetically or evolutionarily unique. I maintain that they are not and I am solely responsible for all the
remaining shortcomings.
2
DeGraff (2003) provides a detailed critique of the “exceptionalism” paradigm.
Journal of Language Contact – THEMA 1 (2007)

www. jlc-journal.org
64 Salikoko S. Mufwene
e.g. Aikhenwald & Dixon 2001; Lightfoot 2002). It only reveals the extent to which languages
share structural materials (sounds, morphemes, or syntactic structures) but not how these came to be
shared.
Adducing arguments from external history, I argue in this article that the above problem stems
from a misinterpretation of the genetic Stammbaum (or cladogram) as an account of language
diversification rather than as a representation of the outcome of the speciation process. The problem
also comes from a basic assumption in genetic linguistics, since the design of the Stammbaum by
August Schleicher in the 19th century, that normal language diversification proceeds monoparentally
(a process identified by biologists as “asexual transmission”). Accordingly, language contact can be
invoked especially to account for irregularities in the correspondences revealed by the comparative
method but not as the cause of the regular changes that lead to speciation. In the case of recent
language contacts, in particular those that have produced creoles, it has usually been stipulated that
the comparative method does not apply. Such language varieties are claimed to be the outcomes,
allegedly unusual, of changes which are externally-motivated. Consequently, these varieties have
typically been disfranchised by most linguists as separate languages that are genetically unrelated to
the European languages from which they have evolved (see, e.g. Thomason & Kaufman 1988;
Thomason 2001, 2002), despite the contrary sentiments of their speakers, as in the case of Gullah in
the United States (Mufwene 1988). To my knowledge, however, the comparative method has never
been applied in such cases (Mufwene 2003a). Posner (1985) and Trask (1996) should not be
dismissed so casually when they claim that French Creoles and Papiamentu are new dialects of
French, in the case of the former group, or new Romance languages.3
I submit that the whole distinction between internally versus externally-motivated change must
have to do with another legacy from the 19th century: the ideology of language purity, which is itself
related to that of race purity. According to this, hybrids, products of race or language mixing, are
less normal, if not simply, abnormal phenomena (Mufwene, to appear). It is therefore not surprising
that creoles and pidgins, as putatively extreme cases of externally-motivated change, have typically
been suggested to be unnatural developments, even by creolists (e.g. McWhorter 2000, 2001, 2005).
The distinction between putatively the “unusual emergence” of creoles and the “normal
evolution” of non-creole varieties must also have to do with a myopic perception of colonization as
a recent phenomenon, correlated only with the dispersal of Europeans around the world since the
Great Explorations of the 15th century. Unfortunately, this position overlooks, or downplays, the
important ecological fact that, for instance, the emergence of the Romance languages has to do with
the Roman Empire, which is a past instance of colonization. The spread of Roman culture
(including Rome’s political, economic, and military systems, as well as its language) entailed
population movements and language contacts. The latter resulted in language shifts and the gradual
prevalence of the colonial language (which won only a Pyrrhic victory) at the expense of the
indigenous ones.
This expansive and replacive evolutionary process is also true of the birth and subsequent
spread, with modification, of Old English as one of the consequences of the colonization of England
by Germanic populations since the 5th century, although it also resulted in the demise of the
continental European languages that the colonists had brought with them. Increased and/or
additional contacts with the indigenous Celtic languages and other continental European languages
(e.g. Old Norse and Norman French) would likewise lead to both the demise of the latter and more
restructuring of English all the way to its modern forms. Many other examples can be cited,
including the colonization of most of sub-Saharan Africa by the Bantu populations (an expansion
which started about 8,000 years ago), the gradual loss of Pygmy and Khoisan languages, and the
3
This position about the genetic classification of French creoles with the Romance languages was assumed
much earlier by Faine (1937), Hall (1950), and Goodman (1964).

Population movements and contacts in language evolution 65
concurrent diversification of the Bantu languages themselves. I submit that it is in fact possible to
account for language diversification around the world as generally motivated by population
movements and contacts, therefore by language contact even in the cases of the so-called
“internally-motivated changes.” This position of course entails acknowledging language and/or
feature competition and selection as a central part of the engine which drives language evolution.
Once we explain, as in Mufwene (2001), that the fundamental and only kind of contact that
triggers language evolution is inter-idiolectal, then the distinction between internally and externally-
motivated change becomes an artificial one, mostly sociological (see also Pargman 2002).4 All
causes of change in any language are external to its structure, lying in the communicative acts of
speakers, such as the accommodations that speakers make to each other in order to be (better)
understood and in the exaptations they make of old materials to convey new ideas. Contrary to the
confusion in Trudgill (2004), mutual accommodations among speakers are part of the continual
process of competition and selection which changes patterns of variation in a speech community,
reducing them in some cases and increasing them in others.5 Kretzschmar & Tamasi (2003) actually
show that languages have long-lasting memories and that variation is seldom obliterated. My
position is that the dynamic fluctuations subsumed by competition and selection are where we
should find the answer to the actuation question addressed by, e.g. Weinreich & al. (1968),
McMahon (1994), Nettle (1999), Labov (2001), and Mufwene (2005a, 2005b).
While it is necessary to situate the actuation of language change in individual speakers, one must
remember that historical and genetic linguistics, the research areas specializing on language
evolution, are about what Mufwene (2001) calls “communal language.” On this view, languages are
extrapolations constructed from idiolects, resembling what biological species are to the organisms
they have been posited to group or class together. The reason for this clarification is that linguists
have typically discussed the lives of languages relative to communities of speakers. With languages
assumed to be shared means of communication within communities, it is only at the communal
level that important cross-idiolectal patterns of evolution can emerge. However, in order to
understand how evolution proceeds and what causes it, we should not lose sight of the
individualities of speakers, although we must situate them in the histories of their communities,
from which they have both inherited a number of patterns and shaped new ones.
This approach to language as practice (constantly in flux), rather than as a static system, is
consistent with Meillet’s (1921, 1929, 1951) and Hagège’s (1993) observation that language
“acquisition” is partly inheritance and partly recreation. It reflects an idea we also find in Lass’s
(1997) phrase “imperfect replication,” which, while accounting for the so-called “internally-
motivated” changes, raise the question of what distinguishes imperfect replication in first-language
4
Campbell (1998: 286-287) supports this distinction, considering as internally-motivated changes those that
are caused by speakers’ physiological and perceptual peculiarities. I am of the opinion that such factors
which shape particular language varieties are external to language itself taken by linguists to be systems.
They just make us aware of the complexity of factors external to language per se that influence its
evolution. They too have contributed to the emergence of creoles.
5
It is absolutely important to conceive of “competition” here relative to a speech, rather than to a language,
community, because the competing elements can also amount to regional and/or social varieties spoken in
a language community. This perspective bears on the subject of language endangerment and loss to which I
return below. In addition, note that the use of the terms competition and selection here need not be
interpreted to suggest that structural features bear some agency in language evolution. As explained in
Mufwene (2003, 2005a, 2005b), they do not, just like genes in population genetics. Competition in
language has to do with the differential way in which variants are weighted by speakers in a speech or
language community, whereas selection refers to the cumulative outcome of speakers’ choices in
individual speech acts, which ultimately produce a particular language variety, distinct from others.

acquisition from Thomason’s (1988, 2001) invocation of “imperfect second language acquisition”
in relation to the development of creoles.
I argue below that all language changes are externally-motivated, in the sense that motivation
for, or the causation of, change is external to language structure, and contact (situated at the inter-
idiolectal level) has always been an important factor causing changes in the “balance of power”
among competing variants. I will deliberately ignore those other important changes that are due to
innovations or exaptations, which are associated with simple inability to replicate the way other
members of the community speak or with new concepts which cannot be expressed with traditional
forms or constructions.6 To accomplish this demonstration, which will be kept at the macro-level
(that of communal languages), we will review part of the history of Europe as a history of
population movements and language contacts. Because of how colonization is implicated in this
history, we will also discuss the differential ways in which European colonial languages have
evolved outside Europe, including the kinds of issues that have recently interested linguists under
the umbrella of “language endangerment and death.” However, it will help to clarify some myths at
the outset of this critical excursus.
2. Some myths and facts about the development of creoles
2.1 Among the facts are the following: Creoles of the Atlantic and Indian Ocean, our heuristic
prototypes for this historic category of languages, have evolved from the contacts of non-European
populations with European populations of principally low socio-economic backgrounds, (native and
nonnative) speakers of nonstandard European vernaculars in exogenous plantation settlement
colonies (Chaudenson 1992, 2001, 2003). These territories, mostly insular and coastal, and typically
between the tropics, were third-party places in which both Europeans and non-Europeans were
newcomers. The colonists developed derivatives of European socio-economic systems that
depended on slave labor, who quickly became the overwhelming demographic majorities.
These societally multilingual communities gradually evolved into monolingual ones in which the
economically and/or politically dominant European language prevailed at the expense of all the
others, an experience that is true of other non-plantation settlement colonies such as North America
– with the exception of coastal Georgia and South Carolina, as well as Louisiana – and Australia.
Language shift among the free populations was experienced later, in the 19th century in North
America, than it was among the slaves and apparently also among the indentured servants. New
colonial varieties of European languages evolved concurrently, in the form of koinés among the
native speakers but in more divergent forms among speakers of different non-European
ethnolinguistic backgrounds, especially after the European and non-European populations became
segregated. In the plantation colonies, primarily those associated with the cultivation of sugarcane,
rice, or coffee, the most divergent vernaculars associated with the slaves and their descendants have
been disfranchised by linguists, under the label creoles, from genetic kinship with the other colonial
varieties that are considered dialects of the same metropolitan languages.7
As much as linguists have denied it, the race of speakers seems to have been an important factor
in this discriminating practice (Mufwene 2001, DeGraff 2003), owing largely to the fact that the
6
But even such changes can spread in the communal language only through inter-idiolectal contacts. In some
communities, especially colonial ones, some of the influential idiolects are xenolectal, associated with L2-
acquisition.
7
The folk term in the relevant communities has been Patois (now nativized as Patwa) in Anglophone and
Francophone territories. It has borne more or less the same meaning as during the colonial period and
before, viz., ‘rural or provincial, especially nonstandard variety of a language, usually derided as bad
and/or unintelligible by speakers of the urban or standard variety’. For discussions of the term creole and
how linguists have extrapolated it, see Chaudenson (1992, 2001, 2003) and Mufwene (1997a), as well the
sources cited in both publications.

distinction between creole and non-creole originates in the 19th century ideology of pure languages.
Subsequent extrapolations of the term creole to language varieties that evolved from colonial
contacts involving no European languages or to others that emerged from contacts of European
languages before the 15th century (e.g. the Romance languages) only illustrate adequately the
perception by some scholars of undeniable similarities between the contact ecologies of the
emergence of creoles and those of the evolution of other languages, as well as between the
mechanisms involved in their restructuring processes.
2.2 Among the myths are the following which have been disputed principally by Chaudenson
(1973-2003) and Mufwene (1996-2005a). Space considerations make it unnecessary to repeat the
arguments here; interested readers may find them in the above references. Suffice it here to expose
and dismiss them very quickly, so that we can reinterpret European language history also from a
colonization perspective.
2.2.1 Creoles developed abruptly, formed by children from erstwhile pidgins made by their parents.
History suggests, instead, a gradual development from the colonial koiné ancestors spoken as
vernaculars by the Creole populations of both European and non-European descent in the
homestead communities that preceded the plantation communities. The intimate, though by no
means equal, living conditions between slaves, indentured servants, and other early colonists made
impossible the development of pidgins as reduced means of communication based on sporadic
contacts. They precluded the possibility that Black and Mulatto Creoles would have spoken
different varieties from White Creoles, especially if population structure rather than biological race
(as controversial as the notion is) determines what particular language variety a child “acquires”
(Mufwene, to appear). Both Creole and Bozal children seem to have played an important role as
transmitters of the colonial language and as an obstacle to a more rampant spread of substrate
influence in the emergent varieties, although, as DeGraff (1999) observes, they must have
contributed to the selection of some of their parents’ xenolectal peculiarities into the new systems.
Moreover, as shown in Mufwene (2005a), there is an interesting geographical complementary
distribution between the territories where creoles developed and those where pidgins emerged, as
illustrated in Map 1.
Map 1: Complementary geographic distribution of creoles and pidgins

2.2.2 It has been claimed that creoles reflect imperfect L2-learning of the European colonial
languages by the slaves (e.g. Thomason & Kaufman 1988, Kaufman 2001). The emergence of other
colonial varieties which are also divergent from the metropolitan ones, e.g. American and
Australian Englishes, Québécois French, or Brazilian Portuguese, suggests that imperfect learning
applies to all groups. No operational yardstick has been proposed to measure the threshold past
which one divergent variety qualifies as a creole and below which it does not. No history-based
characterization has been provided of the unique modes of language “transmission” and
“acquisition” that must be associated with the emergence of creoles. One must look for an
explanation other than imperfect language learning to account for the undeniable structural
differences between the creole and non-creole colonial varieties of European languages we know
today.
2.2.3 Since about Polomé (1983), it has also been claimed that structures of creoles are so divergent
from those of their non-creole kin because there was a break in the transmission of their “lexifier,” a
convenient misnomer for the languages from which they have evolved (Mufwene 2005a). As
McWhorter (2001) puts it, a new language was invented from, or after, the “pulverization” of an
earlier one.
This “discontinuity hypothesis” does not explain why creoles retain such large proportions of
their vocabularies from their lexifiers (up to 95% according to Cassidy (1980), over 85 % in
Saramaccan, one of the most “radical” creoles), nor why we can link so many of their structural
features to their nonstandard European ancestors (see, e.g. Chaudenson (1992, 2001, 2003) and
Corne (1999) for French creoles, as well as Mufwene (2001) for English creoles). In fact, the
position cannot explain why in all plantation settlement colonies the slaves developed their
vernacular from a European, rather than an African, language. Even Berbice Dutch, which has
retained more grammatical materials from a substrate language (Eastern Ijo, per Kouwenberg 1994)
than any other New World creole, has evolved primarily from Dutch. Although Surinamese creoles
have been heavily influenced by Dutch, most of their lexical and structural materials originate in
English, the language of the founder colonists. The Dutch colonists simply decided to communicate
with their slaves in the language that their English predecessors had used, although they themselves
influenced the emergent varieties too, as evidenced by the materials of Dutch origin in Sranan.
2.2.4 It has also been claimed that the comparative method cannot apply to creoles in order to
determine whether they are genetically related to their “lexifiers” (Thomason & Kaufman 1988,
Thomason 2002). The objection to this claim is very simple: it has never been tried and, as pointed
out by Chaudenson (1992, 2001; Mufwene 2003a), the myth is based on mistaken comparisons of
creoles’ structures with those of the standard varieties of European languages. Creoles may actually
reveal the artificiality of the comparative method itself based on “clean” data constituted by written
records, which are not representative of the messy, internally variable, spoken vernaculars. Besides,
as noted in Part 1, the comparative method only shows how much material is shared by a group of
languages, regardless of whether the common stock was inherited from a common ancestor or
spread among them by diffusion. This point is well made by most of the contributors to Aikhenwald
& Dixon (2001). Meillet (1900) had already developed a similar argument, pointing out, in addition,
that genetically related language varieties may share morphological structures or distinctions simply
because they innovated (or borrowed) them under similar ecological conditions, but not necessarily
because they inherited them from their common ancestor. Moreover, we cannot ignore cases where
a language diverges significantly from its genetic kin simply because it has been heavily influenced
by other languages, as in the case of English, which bears heavy influence from Latin, French, and
the insular Celtic languages (among others), compared to Dutch and German.

I argue in Part 3 that creoles are, along with their other colonial kin, the latest linguistic
outcomes of the Indo-European dispersal. The European colonization of the world since the 15th
century, illustrated in Map 2, is just an extension of the original Indo-European dispersal which
began about 5,000-6,000 years ago (see, e.g. Forston 2004) and proceeded according to the cluster-
bomb model illustrated in Map 3.
Map 2: European Colonial Expansion since the 15th century
Map 3: Indo-European Dispersal: The first colonial expansion

2.3 There are a few issues that must be addressed if linguists wish to stick to the assumption that
creoles are peculiar, with their unique set of ecological conditions that set them apart from other
cases of language speciation.
2.3.1 One of these arises from the association of the emergence of creoles with large plantation
colonies, especially those that thrived on sugarcane cultivation. If this association is true of most of
the Caribbean, it is not true of Brazil. Although this polity was the first New World colony to have
launched into sugarcane cultivation and the one with the largest number of slaves engaged in this
particular industry, it has produced no language varieties identified as creoles. The closest thing to a
creole spoken in Brazil is Popular Brazilian Portuguese, which, based on Naro and Scherre (1993),
developed basically from (ex-)colonial nonstandard Portuguese, whose features are clearly traceable
to metropolitan Portuguese. Attempts to classify it as a “semi-creole” (Holm 1989, 2004)
underscore the metropolitan origins of most structural features of virtually any creole, despite the
undeniable role of substrate influence on their restructuring into novel systems. See especially
Chaudenson (1992, 2001, 2003) and Corne (1999) for French creoles.
The case of Brazil need not be confused with that of colonies such as Cuba and Santo Domingo
(the present Dominican Republic). The latter territories had a protracted homestead development
phase, during which their primary industry was animal husbandry, as they gave up sugar cane
cultivation quite early in their colonial history. A large-size and relatively well-integrated Creole
population had already emerged by the late 18th century when they resumed industrial sugarcane
cultivation and imported more and more slaves till the late 19th century. The integration of the bozal
slaves by the large proportion of Creoles prevented the emergence of creole-like varieties, though at
least one significantly divergent variety of Cuban Spanish has been reported, associated especially
with 19th century migrations from Haiti since its independence (Schwegler 2006).
The extent of miscegenation in Brazil seems to reflect a different kind of population structure
which allowed relative integration of the labor force, regardless of whether they were African
slaves, indentured servants, or Native Americans. The case of Cuba and the Dominican Republic
shows that late implementation of segregation could not produce as much divergence as early
adoption of the same practice, such as in the Anglophone and Francophone Caribbean. The case of
Brazil is more comparable to that of the continental USA (except Louisiana), where the late
institutionalization of race segregation in the late 19th century has produced no significant structural
differences between African American vernacular English and White Southern English. The
separation of races came too late to induce a significant speciation of what until then had been the
same vernacular of low-class speakers. (See, for instance, Bailey & Thomas 1998, regarding their
phonological features.)
On the other hand, segregation imposed by geographical isolation accounts for the development
and maintenance of creole vernaculars on islands such as St. Vincent (Prescod 2004) and Cariacou
(Kephart 2000), on which majority populations have always been overwhelmingly of African
descent.8
8
Interestingly, the structures of these creoles appear to be quite close to those of Gullah and thus not as
divergent from their other nonstandard colonial kin as the basilects of Guyanese and Jamaican Creoles, for
instance. This phenomenon raises interesting questions on the progression of the geographical continuum
that Alleyne posited for English creoles as extending from Surinam to North America. Note that the
problem arises only if one assumes, as in the relevant literature on “decreolization,” that the continuum is
linear, with basilectal and acrolectal poles, with partial implicational scales of structural features. I disagree
with that particular view of the continuum, according to which some creoles are allegedly more basilectal
or conservative than others, as if there had ever a risen a common basilect once spoken everywhere in
territories where creoles have evolved from the same “lexifier” (Mufwene 1994, 2001). See also note 10.

2.3.2 Some colonies such as the Netherlands Antilles and Cape Verde have produced varieties
identified as creoles (Papiamentu and Criolou, respectively), although they hardly developed any
noteworthy agricultural industry. For a long time they did not develop a large-size permanent slave
population, as they served primarily as trade posts and slave depots where slaves were in transition
to other destinations. Is it justified to associate the development of creoles with large plantations
and treat the emergence of Popular Brazilian Portuguese as an exception from the rule?
The case of Brazil certainly highlights population structure as a more important factor than the
high demographic disproportion between the native and nonnative speakers in favor of the
nonnative group. However, it also appears that where segregation or isolation of the relevant
population (that engaged in the process of language shift) is a factor,9 we must also take into
account another factor discussed in Mufwene (1996), viz., rapid population replacement in a
steadily growing overall population. This seems to be a critical factor for Cape Verde and the
Netherlands Antilles. Brazil differs from them in developing a less segregated society, although
there are allegations of race discrimination against populations of African descent, as in most other
colonies. Creoles must have evolved as significantly divergent and segregated vernaculars not only
on the plantations but also in those colonies that served as major slave depots. Thus, the
development of creoles need not be associated with one particular kind of economic activity.10
2.3.3 If (extensive) societal multilingualism is a condition for the development of creoles, this factor
does not seem to have mattered in the case of Berbice Dutch, since this variety developed out of the
contact of two major populations: the Dutch and Eastern Ijos. I raised this question in Mufwene
(1993), but I have not seen anybody address it yet. Also, if relexification were really involved in the
development of creoles, this is one particular variety that would prove it. However, even detailed
studies such as Kouwenberg (1994) fail to show the exclusive or predominant origins of Berbice
Dutch’s structures from Eastern Ijo. It is possible to account for its emergence as a by-product (not
outcome, pace Thomason & Kaufman 1988) of language shift.
The case of Berbice Dutch seems comparable to that of the emergence of the Romance
languages from the gradual shift by Southwestern European populations from their Celtic languages
to Vulgar Latin. An important difference lies in the fact that this ethnographic process was
endogenous in the case of the Celtic populations and exogenous in the case of the Eastern Ijos in
Guyana. It is also different from that of foreign workers’ German or French varieties, because even
the “lexifier” was in an exogenous setting in the case of Berbice Dutch and other creole vernaculars.
9
Note that this factor accounts negatively for why African American vernacular English, unlike Gullah, is not
significantly different from American White Southern English. As noted above, segregation in the
hinterland of the American South was institutionalized only in the late 19th century, with the passage of the
Jim Crow laws, almost two and half centuries after (former) slaves and (former) white indentured servants
had lived intimately and developed a common American Southern English. Bailey & Thomas (1998) are so
right in situating the divergence of White Southern and African American vernacular Englishes in the late
19th-early 20th centuries.
10
One must also note that there are other colonies, such as the Madeira and Canaries Islands which also
engaged quite early (in the 15th century) in sugar cane cultivation and slavery but did not produce creoles
either. Size cannot be such an important factor, because of what happened in Cape Verde and other small
islands or archipelagos. The industry did not last long in these colonies (Schwartz 1985), although it is not
clear whether they had population structures similar to that of Brazil. On the other hand, it is noteworthy
that the structures of Cape Verdean Criolou are said to be less divergent from those of Portuguese than
those of São Tomense and Principense in the Bight of Biafra. It has also been pointed out that the tense-
aspect number marking systems of Papiamentu are not exactly like those that are attested in most creoles
(Andersen 1990). Could it be that these creoles developed primarily from the varieties of permanent Creole
populations that spoke what Chaudenson identifies as “approximations” of the European language and
underwent less influence from Bozal speech than elsewhere?

Moreover, no complete shift is involved in these more recent cases. If one considers the fact that
children become gradually monolingual in French or German, then we may assume that the variety
will die soon after the immigration of foreign workers ends.
These differences make the distinction between the development of creoles and that of non-
creole varieties non-structural, pointing to no particular language restructuring process that can be
identified as creolization. Thus, Mufwene’s (2000a) conclusion that creolization is a social process
remains valid. Noteworthy in this connection is also the fact that, as pointed out by Chaudenson
(2001), none of the studies of foreign workers’ xenolectal varieties has produced evidence for
relexification, as homogeneous as the populations that produce them have been ethnolinguistically,
viz., Turkish-speakers in Germany and Arabic-speakers in France. The homogeneity has also been
maintained thanks to residential and/or social segregation, which has left the adult migrants (not
their children) to interact mostly among themselves and to (learn to) speak the local vernacular
occasionally, mostly at work and at the market with non-Turks and non-Arabs, respectively.
If this comparison sheds any light on the development of Berbice Dutch, from an ethnographic
perspective, it suggests that language shift among the Eastern Ijos in Guyana must have been more
gradual than elsewhere, because they could communicate in Ijo among themselves. It also suggests
that a key factor in cases of complete language shift is the “refusal” by children to speak their
parents’ languages, especially in situations where no more speakers of the ancestral languages are
brought to the exogenous setting (Mufwene 2004, 2005a).
2.3.4 Are the so-called “semi-creoles” evidence of a process of “creolization”? Or do they suggest a
continuum of degrees of structural divergence from the “lexifier”? In the latter case, they may be
interpreted to suggest that, subject to specific ecological conditions, the restructuring process did
not always yield varieties that are equally divergent. At the communal language level, one can say
that some varieties remain closer to the terminus a quo than some others.11
The best empirical answer to the latter question comes from Schneider’s (1990) study of English
colonial varieties spoken by the descendant of Africans in the Caribbean and the United States. He
presents a “cline” of divergence from standard English, showing African American vernacular
English (AAVE) to be less divergent than basilectal creole varieties of especially Surinam and
Guyana. The study corroborates in quantitative terms what Alleyne (1980) had already observed by
positing, in addition to local social continua, a broad geographical continuum of “Afro-American”
extending from Saramaccan in Surinam to AAVE in the United States that he thought was a
concurrent of the differential development of creoles rather than the outcome of decreolization as
claimed by Schuchardt (1914), Bloomfield (1933), DeCamp (1971) and many later linguists who
assume Hall’s (1962) “life-cycle” theory.
As shown by Lalla & D’Costa (1990) and argued in Mufwene (1991, 1994), there is no
convincing diachronic evidence (perhaps with the exception of Barbados, per Rickford & Handler
1994), for decreolization in creole societies, let alone for the evolution of AAVE from an erstwhile
11
As explained in Mufwene (2001, 2005a), there cannot really be a common yardstick for measuring degrees
of divergence, because the “lexifier” was hardly the same from one colony to another. Assuming that the
European languages were internally variable, the different colonies hardly constituted identical
representations of metropolitan variation patterns, especially in terms of the demographic strengths of the
different variants. From an evolutionary perspective, this cross-setting variation entailed different
dynamics of coexistence, different markedness values associated with the variants, and thus the possibility
that different variants could be favored in different settings. As in much of the literature, the present
discussion oversimplifies a colonial situation in which metropolitan languages evolved differentially from
one setting to another. Besides, the evolutions were primarily local (Kretzschmar 2002); one must indeed
remember that polities are in effect “metapopulations” of localities interconnected by a transportation
infrastructure (traditionally roads and rivers) and by “dispersing individuals” (Mufwene 2001).

Gullah-like English that would have been spoken all over the USA in the 18th and/or 19th centuries
(Mufwene 1994). Recent scholarship, such as Labov & Harris (1986), Bailey & Maynor (1989), and
Bailey & Thomas (1998) suggest divergence evolution between AAVE and American White
Southern English, just like Pollard (2000) suggests non-convergent evolution between Jamaican
Creole and Jamaican English.
The greatest problem with the semi-creolization hypothesis as the reversal of the decreolization
hypothesis lies in difficulties with determining what “creole features” really are or what a
“prototypical creole” is supposed to be. McWhorter’s (1998) attempt at identifying the prototypes
by the combined absence of inflections, of derivations, and of tones was not particularly successful.
With extensive counter-arguments, DeGraff (2001) demonstrated that the alleged lack of
derivations was no more than a controversial parameter in the case of Haitian Creole. We may want
to add that the “lexifiers” are not particularly tonal languages, which can naturally explain why
creoles need not be expected to be tonal, although most of their substrate languages are. The
competition of features was naturally resolved in favor of toneless systems, except in Papiamentu.
In addition, the nonstandard varieties from which creoles have evolved are not as inflectionally rich
and regular as their standard varieties (Chaudenson 1992, 2001). The rarity of inflections in creoles
is thus, as Chaudenson (ibid.) already observes, the ultimate conclusion of an inflection-erosion
process that was already taking place in the colonial “lexifiers” themselves. Publications such as
Brasseur (1997) and Chaudenson et al. (1993) make this observation more evident.
Coming back to derivations, one can only thank DeGraff (2001) for showing that creoles have
not only selected many affixes from their “lexifiers” but also innovated other ones, despite the loss
of some and the significance of multifunctionality in their morphosyntaxes. Consistent with the fact
that the Romance and Germanic lexifiers themselves represent impoverished morphologies
compared to their older Latin and Germanic ancestors, respectively, we can safely assume, after
Chaudenson (1979f), that the so-called “creolization” process is largely an extension of
restructuring processes that had already started in the relevant “lexifiers.”
There cannot be a “creolization” processes without a specific combination of “creole features”
that would help us identify it. There is no exact measure that can be used to determine whether the
restructuring has proceeded only half, or part of, the way. Rather, the reality in which every creole
differs from every other creole that evolved from the “same” colonial European language shows
family resemblance, in the same way that the Romance, West Germanic, or Bantu languages share,
within their respective families, only parts of their structures with each other, while remaining
different in a number of other respects.
The “semi-creolization” hypothesis is thus an inadequate way of capturing the differential
evolution of colonial European languages in plantation settlement colonies, with some of the new
varieties being less divergent from their colonial non-creole kin than others. It reflects a
misinterpretation of the ways in which differing ecological settings (e.g. proportions of speakers,
changing strengths of the variants in the super- and substrate languages, patterns of population
growth) have variably influenced the outcomes of the contacts of more or less the same languages.
The remark applies to observations such as in Winford (1997, among others), which identify some
creoles as “intermediate” or “mesolectal” (e.g. Bajan), as if a uniform basilect was once spoken
everywhere in the Caribbean (and North America) by all slaves. Alleyne’s (1980) position on social
and geographical continua remains the one most consistent with history.
This ‘ecological setting’ is also made more complex by Chaudenson’s (2003) parler de seconde
genération,12 which is also evident in Le Page & Tabouret-Keller (1985) and Winford (1997), in
12
This is an improvement over what he originally misformulated as “generations of creoles” on the models of
generations of computers in Chaudenson (1979, 1992, 2001). The colonial language variety brought to a

reference to varieties such as Guyanese Creole or Mauritian, whose evolution cannot be explained
without factoring in the colonial language variety imported to the new colony from an earlier
colony. Colonial history suggests that cross-colony influences must have been more common: for
instance, in the case of Anglophone Caribbean, from St. Kitts to Barbados and other islands in the
early 17th century (Baker & Bruyn 1998), from Barbados to Surinam and Jamaica in the mid-17th
century, from Barbados to South Carolina in the late 17th century, from South Carolina to Georgia
in the early 18th century, and from Barbados and other islands to Trinidad and Tobago and to
Guyana in the 18th century. As suggested by Le Page & Tabouret-Keller (1985), this complex web
of colonial migrations defies a unilinear account of the evolution of creoles.
2.3.5 The above discussions must prompt us to ask whether, in the first place, it is rewarding, from
the point of view of language evolution, to assume a priori that creoles have developed in their own
unusual way. While it is informative to find out the respects in which they diverge structurally from
(the standard varieties of) their “lexifiers” and their other colonial, non-creole kin, how important is
it to label them seemingly a priori as creoles? For instance, from an evolutionary point of view, how
much is to be gained from identifying the Cape Verdean vernacular as creole when comparing it
with Popular Brazilian Portuguese? The same question applies to Gullah, also identified as a creole,
in comparison with its American hinterland kin AAVE, which is identified by Holm (1989, 2004)
as semi-creole. Even the identification of the latter as a semi-creole is question-begging in
comparison with other colonial varieties such as Old Amish English, which has clearly arisen from
the contact of English with some continental European languages (notably German), under
particular conditions of social isolation, but is not identified as a (semi-)creole.
Perhaps the extent-of-divergence question, arising from the term semi-creole, is revealing from a
sociohistorical perspective. Insofar as the identification of particular colonial varieties as creoles
can generally be correlated with places where descendants of Africans have become overwhelming
population majorities, the term presupposes some measure of the extent to which the non-Europeans
were socially isolated and of whether miscegenation has been extensive or limited. From this
perspective, even varieties such as Cape Verdean are creoles, although its structures are said to be
(much) closer to nonstandard Portuguese than those of São Tomense, Angolar, and Principense (in
the Bight of Biafra). The particular ecological conditions of their development, including patterns of
population growth, the ensuing population structure, and the proportion of speakers of
(approximations of) the colonial European language among non-Europeans can help us account for
the differential outcomes of the structures of specific creoles.
However, what we are not learning is the more general fact that, regardless of whether or not the
new varieties are identified as creoles, the specific local ecological conditions of population
contacts generally influenced the differential evolution of the European languages from one colony
to another. For instance, the French variety of Québec differs from that of Louisiana just as
Martiniquais Créole does from Haïtien, because of specificities of the ecologies of their emergence.
We can explain in the same way differences between, for instance, American Southern English,
New England English, Gullah, AAVE, Jamaican Creole, and Jamaican English. One important
reason why the traditional approach to the development of creoles should be given up is that it has
generally failed to inform us, from the perspective of language restructuring, about the specific
colony from an earlier colony need not have been a creole already. In some cases, such as when Mauritius
was settled from Réunion or Suriname and Jamaica from Barbados, it is doubtful that a creole had already
emerged in the mother colony. What matters the most in this particular case is the fact that second-
generation colonies did not start from the same kind of language contact situation as their first-generation
counterparts. The slave populations from the earlier colonies already spoke some variety of the colonial
language and provided a more coherent, or less diffuse, linguistic infrastructure from which the creole
would develop, perhaps faster.

respects in which the emergence of creoles generally differs from the presumably more continuous
evolution of, for instance, the Germanic and Romance languages from their proto-varieties.
As in Mufwene (2001, 2005a), I argue below that external history suggests that there are actually
no differences in regard to the restructuring processes involved. We could as well have spoken all
along of the continuous, gradual evolution of creoles from their “lexifiers.” If anything, creoles are
prompting us loud and clear to question some of our assumptions about the putative differences
between their development and those of the non-creole languages, especially regarding the critical
role of language or dialect contact in answering the “actuation question.” Variation in the specific
language varieties involved in the contacts as well as in population structure, along with other
ecological factors discussed in Mufwene (2001, 2005a, 2005b), go a long way to explaining
language speciation, highlighting similarities between the evolution of both creoles and non-creole
language varieties.
Those like Thomason (2002, 2003) who are particularly wedded to the comparative method as
proof of genetic kinship among languages should also remember that all shared items and structures
among languages are not necessarily due to inheritance from a common ancestor; contact can also
account for some of the formal and structural correspondences (Aikhenwald & Dixon 2001, Heine
& Kuteva 2005) as can parallel innovations (Meillet 1900). A contact-based approach to genetic
linguistics can thus be informative not only about why some offspring of particular proto-languages
have more divergent structures than others, but also about how the divergence can be correlated
with patterns of other languages that a proto-language came in contact with. The approach can thus
inform us about shared forms and/or structures that are not attested in the proto-language. These are
precisely the kinds of explanations that have been provided for why creoles have, for instance, well-
developed serial verb constructions, due to the influence of substrate languages that have such
constructions. All things considered, the uniformitarianism advocated in this essay makes it also
natural to invoke inheritance of features from the lexifier as an explanation for why, for instance,
English creoles have prenominal definite articles and attributive adjectives as well as stranded
prepositions in questions and relative clauses, unlike their substrate languages. Below, I start an
excursus, based on external language history, which supports this uniformitarian thesis.
3. The “creole”-like evolution of English and the Romance languages
Similarities between the development of, on the one hand, creoles and, on the other, English and the
Romance languages have been pointed out before in creolistics. Bailey & Maroldt (1977)
hypothesized that Middle English must have creole origins, heavily influenced as it was by Norman
French. They actually overlooked the fact that little, or no, language shift was involved in this
particular case among the Germanic populations in England. The majority of English people
continued to speak their vernaculars, little influenced by this continental European language. Only
one sociolect, heavily influenced by the speech habits of the Norman colonists (aristocrats and
administrators) and used then by their English colonial auxiliaries (a small minority), has evolved to
show most of this French influence. It is standard English, used here as an abstraction of
convenience subsuming several national standards. It is the sociolect that has undergone the Great
Vowel Shift in its fullest form, representing an interesting case of differential evolution in a
communal language. This was happening at a time when the more indigenous populations, the Celts
(rather than the Germanics who had colonized them since the 5th century), were gradually shifting to
English and seem to have been influencing the evolution of its vernacular varieties, as now made
more obvious by students of Celtic Englishes (see, e.g. Filppula & al., eds. 2002). The completion
of the shift is marked more conspicuously by the emergence of vernaculars such as Irish English.
In fact, it would have been more interesting to approach the genesis of English itself as a
colonial, contact-based evolution that affected the Germanic language varieties brought to the
British Isles, as noted in Part 1. Though it is not clear whether there was a single “lexifier” in this

particular case, it is very interesting that one particular group, the Angles, bequeathed their name to
both the land now known as England and the language, English, which can now hardly be defined
as ‘the language of the English people’ (Mufwene 2000b). While ecologically its genesis is
apparently comparable to koinéization in the emergence of colonial Englishes, its subsequent
evolution in the British Isles is very much influenced by important changes in the English
population structure and by a continuous process of language shift among the colonized Celts,
making the phenomenon no less interesting than that of later colonial English varieties, including
indigenized Englishes and English creoles (and why not also pidgins).
We need not of course repeat Bailey & Maroldt’s mistake in invoking a putative process of
“creolization,” whose meaning as a structural process remains illusive (Mufwene 2000a, 2001), not
any more than we must disfranchise creoles a priori as having developed in their own unusual way.
We should first abandon the 19th-century myth of language purity and approach the emergence of
all of the new varieties from the point of view of language contact. The next step is then to let the
facts inform us about more adequate distinctions that can be made.
Like Bailey & Maroldt, Schlieben-Lange (1977) compared, in the same volume, the emergence
of Romance languages with that of creoles. If there really were creolization as a global language-
restructuring process that had the peculiarity of transforming a non-creole language into a creole
variety, her comparison was perhaps even closer to the being accurate than Bailey & Maroldt’s. It
was after the Romans had abandoned their southwestern European colonies that the masses of the
continental Celtic populations gradually shifted to Latin, which had become the vernacular of the
local aristocracy and a major trade language. Christian missions and urban centers played an
important role in spreading it. The scenario is similar to that of the development of creoles, after the
European and non-European populations were segregated and mostly Black Creoles and seasoned
slaves served as model speakers for the bozal slaves. See below.
On the other hand, as noted by De Landa (2000), this history of Latin also suggests that it did not
spread in southwestern Europe in a rectilinear fashion. On the contrary, it spread from secondary
dispersal points in the colonies: the Christian missions and the trade and/or administration points
where it was already entrenching itself as a vernacular or as a dominant lingua franca. Noteworthy
also is the fact that it is the nonstandard variety or varieties, Vulgar Latin, which, as accepted
among Romanists, spread within the masses of the indigenous populations and subsequently
evolved into the Romance languages, under Celtic substrate influence.
In this case too, we can note a bifurcated evolution. On the one hand, the elite variety known as
Classical Latin, remained contained in the school system and among scholars, until it was replaced
by the standardized varieties of the new national vernaculars (especially, French, Spanish, and
Portuguese). On the other hand, Vulgar Latin continued its natural evolution into the nonstandard
and colloquial Romance vernaculars spoken today. From the point of view of language vitality, one
can also tell that it is not always the privileged or the most prestigious varieties that thrive. Classical
Latin, is virtually dead, spoken today only as a lingua franca at the Vatican, while Vulgar Latin may
be claimed to still be alive, in mutated forms.
The Celtic experience of language shift to Latin and its restructuring into the Romance
languages is largely reminiscent of language shift and restructuring among the African slaves in the
New World and Indian Ocean colonies of the 17th-19th centuries. Just like the Celts learned Latin
primarily from non-Romans, most of the slaves had non-Europeans as their model speakers,
especially during the plantation phase, when they became majorities in their respective colonies and
segregation was then institutionalized to protect European minorities against possible uprisings. By
the late 18th century, fewer and fewer bozal slaves learned the colonial vernacular from those
Creoles who were likely to speak closer approximations of European speech; they had seasoned
slaves, L2-speakers of the emergent creole, as their models. As with today’s indigenized varieties of

colonial European languages, the more speakers they gained, the more divergent they became,
thanks to the mutual reinforcement of xenolectal features among nonnative speakers. The process
was just as true during the gradual divergence of the Romance languages from (Vulgar) Latin. From
the point of view of language shift, pressure to communicate in the new language came from people
of one’s own kind: urban and Christian missions’ Celts in today’s Romance countries, and the
Creoles and seasoned slaves on the plantations. Ecology-specific considerations of adaptation costs
and benefits to speakers drove the massive shift to monolingualism in the colonial language.13
Although the two cases of language shift differ in that the genesis of the Romance languages was
endogenous and that of creoles exogenous, both kinds of settings favored substrate influence as
acknowledged by Corne (1999), Chaudenson (2001, 2003), and Mufwene (2001, 2005a, 2005b),
i.e., primarily as factors influencing the selection of particular structural variants from among
various competing alternatives in the “lexifier.” In both cases, substrate influence was facilitated by
the fact that the new speakers, those who had shifted, or were shifting, languages spoke the new
vernacular more among themselves than with its native or original speakers. We can indeed observe
this phenomenon now in the so-called indigenized Englishes and les français africains, which
display idiosyncrasies that can be related to features of their dominant substrate languages.
Recall that the term substratum, so central in studies of the development of creoles, came from
Romance linguistics (Goodman 1993), in which its meaning was closer to its application in geology
to lower layers of soil stratification. Having preceded Latin in the now Romance countries, the
Celtic languages represented the “substratum” and Latin the “superstratum,” in a way comparable to
geological layers, which are also chronological. As pointed out by Chaudenson (1990) and
Goodman (1993), Hall (1962, 1966) had extended the terms incorrectly, substituting a social
stratification for the chronological order of arrivals in the colonies.
In many cases, European colonies that developed creoles or other divergent varieties associated
with non-Europeans (e.g. Virginia, Barbados, Bermuda, and Réunion) had been settled by
Europeans before substantial numbers of African slaves were brought in. In others, Europeans and
non-Europeans arrived literally at the same time (e.g. Jamaica, South Carolina, and Mauritius), as
these “second-generation colonies” were being settled from earlier, “first-generation colonies” (in
the language of Chaudenson 1979-2003). There are no exogenous settlement colonies that had been
settled by Africans before the Europeans arrived. In fact, some of the islands, such as Cape Verde,
São Tomé, Principe, Réunion, and Mauritius, had been uninhabited before the European
colonization.14
13
t is important to emphasize here that the experiences of language shift in Hawaii, the New World, and the
Indian Ocean were not identical. Because Hawaii’s contract laborers were ethnically segregated, their
progenitors have maintained their ancestral identities to date. Therefore Japanese, Chinese varieties,
Korean, and Pilipino have survived as ethnic languages. Only the European contract laborers have blended
to some extent with the White Americans and have lost their languages. On the other hand, the slaves of
the New World and Indian Ocean were mixed since the initial, homestead phase of the colonies. The
concurrent emergence of a Creole population that never spoke a pidgin or a creole and served as model to
the bozal slaves contributed to the continuous obliteration of ethnic distinctions, even during the post-
Emancipation importation of contract laborers from regions of Africa and India that were relatively
homogeneous linguistically (Mufwene 2004, 2005a). This is another interesting manifestation of the
Founder Principle, with some of the practices of the founder population having far-reaching consequences.
14
However, as Hookoomsing (2000) points out, in the case of Mauritius, this is not so true of French
colonization. The Dutch had discovered the Island in 1598, colonized it for the first time in 1638 and then
finally abandoned it in 1710 to pirates (Encyclopedia Britannica, electronic edition, 2000), leaving some
Maroon slaves behind (Hookoomsing 2000: 149). It is not clear what language varieties those Maroons
spoke, if they survived the pirates and/or the Island’s colonization by the French in 1721. If they did, what

Nonetheless, creolists have more or less accepted and normalized this notional inaccuracy, i.e.,
the incorrect way the term substratum was transferred from Romanistics by Hall (1962, 1966). They
have capitalized on the fact that African languages, ranked socially as low, influenced the gradual
evolution of some colonial varieties of European languages into creoles, just like the Celtic
languages, also ranked as low varieties, played a role in the gradual transformation of Latin into the
Romance languages. In both cases the influence was exerted through populations that were shifting
from their ancestral vernaculars to a new one, the Celts in the case of Vulgar Latin and the African
slaves in the case of colonial varieties of European languages.
The historical facts in themselves, those having to do with population movements and language
contacts, underscore the need to approach the evolution of Latin and European languages in the
colonies from the same perspective of language competition and selection in varying ecological
conditions, as explained in Mufwene (2001, 2003b). We must simply bear in mind that competition
and selection actually operate at the level of features, with restructuring cum feature-recombination
(analogous to gene-recombination in biology) favoring those from a particular language (misnamed
lexifier in creolistics) but making allowance for features from other languages in contact
(misidentified as substrate) to get in or simply bear on the selection of variants from the
advantageous language.15 In the development of both creoles and the Romance languages, we must
be careful not to confuse the question of the origins of features (lexical items, morphemes,
grammatical rules, etc.), which lie predominantly in the so-called lexifier and sometimes in the
“substrate” languages, with that of influence that the latter have exerted on the selection of
particular variants from a range of competing alternatives in the advantageous language.
Many more ecological similarities emerge from this comparison of the development of creoles
and the Romance languages. They explain why both Posner (1985) and Trask (1996) are probably
not mistaken in claiming that Haitian Creole and Papiamentu, presumably like other creoles lexified
by French and Portuguese, are the latest Romance language varieties to have emerged. I return to
this issue below. Recall, however, that the Latin variety appropriated and modified by the Celts was
the nonstandard variety spoken by the (former) Roman legionaries and low-level colonial
administrators, most of whom had been recruited locally. These are indeed the Latin speakers with
whom the Celts interacted. It is therefore natural that the Latin which they modified gradually into
Old Romance varieties (Old French and Old Iberian, in particular) developed from Vulgar Latin
rather than from Classical Latin.
Also, contrary to what is often suggested in the literature referring to the older age of the
Romance languages, it is actually Old Romance varieties that developed directly from Vulgar Latin,
not the modern Romance languages. The latter represent later, post-formative stages in their
evolution. Arguments that have dodged the comparison with creoles by claiming that it took the
Romance languages over one thousand years to evolve (to their current structures) have missed the
point about their initial formation. It did not take Old Romance much longer that it took creoles to
emerge, i.e., be disfranchised, as separate languages, viz., more or less two centuries from the
impact did they exert on the development of the French colony and of the new colonial vernacular that was
to become Mauritian Creole?
15
I submit that languages are quite osmotic (see also Chaudenson 1992, 2001) and their boundaries are less
rigidly defined than linguists have suggested. Although some institutions try to erect official “language
borders” and impose themselves as immigration and/or customs officers (as in the case of the Académie
Française), the average speaker just ignores these artifacts and feels free to let xenolectal elements into
their language variety. Even if they subscribe ideologically to a particular language they intend to speak,
the evidence of varieties derided as franglais or Spanglish suggests that practice holds “open borders” and
is more open to cross-language “migrations.” Such varieties represent, in the extreme, how competition
and selection work in multilingual communities. Needless to say that there are ecologies, such as those
where creoles developed, which strongly favor one particular language.

adoption of the language by the masses of the Celtic populations. If anything, we have no idea what
the structures of creoles that survive will be like another thousand years from today. Nor can we
surmise whether they will become so autonomous from their lexifiers as to develop their own
separate standard varieties.
In any case, there is no doubt that the same kinds of shift and restructuring processes have been
involved in the evolution of the Romance creoles and that of the Romance languages and their
recent non-creole offspring. Specific studies of the respective restructuring of Vulgar Latin and
colonial European languages into new varieties can only help us articulate the specifics of this
approach more concretely, consistent with the ecologies in which the varieties have emerged. I will
now show why it would not be far-fetched to argue that creoles are among the latest Indo-European
vernaculars to have emerged.
4. The linguistic consequences of the latest Indo-European Expansion
The foregoing discussion can be interpreted in the following ways. The dispersal of the Indo-
Europeans from their homeland must have proceeded in more or less the same way as the settlement
colonization of especially the New World, Australia, New Zealand, and Falkland Islands, not to
mention the Atlantic and Indian Ocean islands around Africa, since the 15th century. The original
populations need not have been politically and ethnolinguistically homogeneous, as well argued by
Trubetzkoy (1939). They need not have departed from exactly the same geographical location
either. Regardless of whether this lay in Asia Minor or Caucasia, it was a broad area. Neither need
they have left the homeland at the same time, not any more than they could have taken the same
dispersal routes. Nor did they reach their destinations at the same time. The histories of, for
example, the Hellenic the Roman Empires, as well as that of the Germanic expansion westward and
southward from Scandinavia, all suggest also that the original colonization routes led them to new
dispersal points from which they spread in all sorts of directions. This dispersal must have occurred
in a way that is not matched at all by the traditional cladograms of genetic linguistics, with
rectilinear and non-intersecting distribution branches, even if they are adapted as in Map 4 to match
the present geographical distribution of IE language families.
Map 4: A Geographically-Adapted Indo-European Cladogram

Colonial expansions, often associated with mercantilism (Cowen 2001), were hardly ever
planned. They were not at all like military expeditions, orchestrated by a central senior officer,
although many of these were used to support them. For instance, no particular army-like general
directed the Celts to be among the first to move out of the IE homeland and to wait for the
Hellenics, the Italics, and the Germanics to come and colonize them, at different times, in those
territories where they (the Celts) had settled much earlier, as illustrated by Map 5.
Map 5: The Non-Rectilinear Expansion of the Indo-Europeans (same as Map 3)

The history of the Indo-European expansion in Europe is one of population movements,
therefore of language contacts, not only with the pre-Indo-European populations that had inhabited
the territories they came to settle or control but also among themselves. As clearly shown by Heine
& Kuteva (2005), one cannot account for the diversification of Indo-European languages and for
their shared structures without factoring in language contact. This paper shows that the rationale for
approaching the subject matter as I just explained is the same as for accounting for the speciation of
some Western European languages into creoles and for how they variably share only some of their
Indo-European structural features but not others, in ways that can be correlated with their contact
histories.
We could thus also conclude that the recent colonization of the world since the 15th century is
but an extension of the same Indo-European expansion that started 5,000-6,000 years ago. As in the
case of the colonization of England by the Germanics, the recent colonization of the world by
Europe started with small expeditions of pioneer colonists, who would be joined later by larger
cohorts of immigrants who imposed their socio-economic systems, although these were adapted to
local physical ecologies. In settlement colonies, they became majority populations, with the
exception of plantation settlement colonies, in which the non-European labor populations (slaves or
contract laborers) became majorities especially over the indigenous populations.16
16
As explained in Mufwene (2004, 2005a), the Roman colonization was not really on the settlement model. It
was perhaps closer to the exploitation model that applied to Asia and Africa, although the Romans were

The particular experience of settlement colonies is interesting because it entailed language shift
in favor of (one of) the exogenous ruling populations. However, the latter’s languages won only
Pyrrhic victories, modified as they were by the languages that they displaced. Just as Latin was
restructured into the Romance languages, most of the West Germanic languages taken to England
died (like several European languages in North America), being displaced by the language of the
Angles, which immediately speciated into several British varieties and would continue to diversify
into new ones outside the British Isles. Overall, in every extra-European settlement colony, the
dominant European language has displaced not only indigenous languages but also other European
languages with which it competed.
Scholars opposed to my perspective will want to capitalize on the fact that although they have
evolved from the same colonial European koinés (Chaudenson 1992, 2001, 2003, Mufwene 2001,
2005a), creoles are structurally more divergent from their European metropolitan sources than their
non-creole colonial kin (e.g. North American English varieties). In the first place, the reliability of
such judgments depends on the kinds of varieties being compared. It is not so clear that structural
differences are so strong between Cajun French, Louisiana French Creole, and Louisiana French, or
between Gullah and Old Amish English, especially if one overlooks prosodic features. Be that as it
may, the current literature has usually omitted to consider population structure as an ecological
factor that can account for structural differences between the colonial vernaculars spoken by
descendants of Europeans and those spoken by descendants of African slaves especially in the case
of North America.
Like the indentured servants of the pre-independence period (e.g. the period preceding the
American Revolution), the slaves shifted early to their masters’ languages. As soon as they became
the plantations’ dominant populations, segregation was instituted, quite early in the 18th century,
forcing them to interact and socialize more among themselves than with populations of European
descent. As the proportion of the Bozal slaves kept increasing, more and more Bozals learned the
European colonial vernacular from seasoned slaves, whose command of the language was already
divergent, than from the original Black Creole speakers, whose competence differed little, if at all,
from that of their White Creole counterparts. During this continual process of language shift, the
colonial vernacular underwent more and more substrate influence and diverged increasingly to
reflect this substrate influence, bearing in mind that the influences themselves were subject to
competition and selection among the slaves (Mufwene 2001, 2003b).17
This situation is quite comparable to that of the gradual shift of the continental Celts to Latin as
their vernacular. Being the majority populations increasingly communicating among themselves in
the new, adopted vernacular, their Celtic linguistic habits apparently influenced the restructuring of
Vulgar Latin into today’s Romance languages, which are as different from Classical Latin as
creoles are from the standard varieties of their “lexifiers.”
There are very good ecological reasons why the colonial varieties spoken by descendants of
Europeans diverge less from the metropolitan varieties than creoles do, assuming correctly, like
more interested in benefiting economically from the imperial enterprise than in claiming political and
economic control rights over every square inch of every colony. Rather, they developed networks of city
colonies administered by Romanized indigenous aristocrats, whom they rewarded well. This all shows that
language dispersal does not proceed everywhere through a single population structure, suggesting that even
today’s pidgin and indigenized varieties of European languages may very well count as new offspring of
Indo-European languages.
17
None of what I say here need be construed as a change of my position from Mufwene (2001, 2005a). Recall
that influence need not be confused with origins of features. The nonstandard varieties of European
languages that became colonial vernaculars were internally variable. The substrate languages influenced
the selection of particular features, especially favoring those that were (partially) congruent with those
already familiar to the learners.

Chaudenson (1992, 2001, 2003), that the divergence is generally a later development. As also
observed by Chaudenson, most of the nonstandard vernaculars of the European languages that came
in contact with each other are typologically more similar among themselves (as made more obvious
by Heine & Kuteva’s 2005 demonstration that Europe is an important linguistic area marked by
convergence) than they are with the African languages. Equally important is also the fact that the
European populations remained largely segregated by nationality until the early 20th century. Until
then, they kept their national vernaculars and used the economically (and politically) dominant
language only as a lingua franca.
Thus, the colonial English varieties now spoken by descendants of Europeans seem to have
developed primarily among colonists from the British Isles, with only minimal “adstrate”
influence18 from the other European languages that they displaced later. At the time of the shift in
the late 19th and early 20th centuries, the adults must have gradually died out with their accents, just
like many adult immigrants do today, while their children, through regular interactions with the
children of native speakers (especially in school), acquired the dominant vernacular natively. This
process, which can still be observed today among the non-Aboriginal populations of Australia
(Clyne 2003), has kept adstrate influence to a minimum. In fact, it is still true of recent and current
immigrant families in North America, as of anywhere else in the world.
The above explains why, in North America, it has become less and less common to speak of, for
instance, German and Italian Englishes. The populations of English and non-English descents have
increasingly blended to the point of reducing ethnic differences among them, at least among those
who are not first-generation immigrants. Most of the linguistic peculiarities that justified those
ethnic labels have been obliterated, while a few of them may have spread in general American
English. Only varieties of the communities that have still not been integrated in the dominant
population still carry such ethnic labels (e.g. Old Amish English) in some way comparable to
regional labels such as Ozark and Appalachian Englishes. Yiddish English, associated with Jewish
first-generation immigrants from Yiddish-speaking territories is considered moribund.
Interesting from this perspective is also the fact that 20th century African and Caribbean
immigrants to North America have exerted no noticeable influence on the structures of African
American English, largely because they have generally not settled in African American
neighborhoods and have socialized with them in no more significant ways than they have with other
Americans. When this has been the case, these immigrants’ children have learned to speak AAVE
natively. Caribbean immigrants who have settled in communities in which they either are dominant
or constitute a critical mass have maintained their native Caribbean features, whereas their children
have had to select between their parents’ and American English features.19 The same observation
applies to any ethnic community, unless it has been formed recently and just reached its critical
mass.
Innovations generated by the internal ecology (i.e. the dynamics of both intra- and inter-
idiolectal variation within a communal system) continue of course, such as in the Valley Girl Talk
in California, Hip Hop language among American teenagers, and Dread Talk in Jamaica. However,
18
My use of the term adstrate here is just an unfortunate legacy of creolistics that could be done away with. It
continues to suggest that substrate influence is from populations that are ranked lower while adstrate
influence is from populations that are of more or less the same social stratum. Linguistically, the influences
do not work in different ways, especially in the recent European colonial settings since the 15th century.
19
Although the phenomenon has emerged in another polity, London West Indian Patwa, which is not spoken
by all offspring of Caribbean immigrants (even in those cases where the parents are Patwa-speakers), is
just an instantiation of this selective process. London West Indian Patwa is in fact spoken primarily by
those immigrants who do not feel integrated in the British social structure and use language as an identity
marker.

very little of this innovative speech has been motivated by contact with the languages of the recent
immigrants. Inter-group variation within national varieties also persists, because it is determined by
patterns of social interaction; schooling and the mass media have little bearing on it. Labov’s (2001)
discussion of the Northern Cities Vowel Shift also reveals another interesting aspect of how
changes spread: only among people who interact with each other, in this case urban white
populations, a reflection of the persistent ethnic segregation in the American population structure.
All these differential evolutions underscore the significance of population structure (Mufwene
2005b) as an ecological factor. It determines not only who interacts with whom but also who
accommodates whom and under what ethnographic circumstances.
This paper also makes more evident the fact that language shift, which has recently been
associated with the demise of the ancestral languages of several indigenous and immigrant groups
in European settlement colonies, has been a concomitant of the spread and diversification of Indo-
European languages. From a longer perspective, this evolution has been in process since the first
territorial expansion of Proto-Indo-Europeans in Asia and in Europe to the latest dispersal that has
produced new colonial offspring of some of the modern Indo-European languages. Extinct
languages have left their marks (substrate or adstrate influence) on those that they displaced them.
A more accurate account of language vitality in genetic linguistics should thus provide a balance
sheet of losses and gains, instead of capitalizing exclusively on losses and bemoaning loss of
diversity and ignoring the emergence of new patterns of diversity (Mufwene 2003c, 2004, 2005a).
Interestingly, very few of the massive shifts from indigenous vernaculars to European languages
have been observed in European exploitations colonies of Asia and Africa,20 where the same
languages have been adopted as official languages and as elite lingua francas. Despite the fact that
the latter have indigenized, as evidenced by the growing literature on indigenized Englishes (see,
for instance, the references in Bolton 2002, Kachru 2005) and “les français africains” (e.g. Lafage
1977, Manessy 1994, Manessy & Wald 1984), rare are the cases where Native Asian and African
vernaculars are being driven out by the colonial European languages of their respective polities.
That Colored people in South Africa have appropriated Afrikaans as their vernacular is more a
consequence of the dual colonization system in this polity, with the Afrikaners’ socio-economic
domination representing settlement colonization and the British the exploitation model. However,
recent changing political tides are increasingly reverting the ethnographic status of Afrikaans to that
of an ethnic language that is economically less rewarding, causing more and more Colored people
to raise their children as native English speakers. Afrikaans may become one of the rare endangered
languages with millions of speakers, reminding us that it is not so much the number of speakers in
abstraction that determine the vitality of a language but how it negotiates speakers with the other
languages it competes with and whether the proportion of children acquiring it as a vernacular
increases or decreases.
The explanation of this differential evolution of colonial languages is largely ethnographic. In
Asia and Africa, the colonial languages have typically been appropriated as lingua francas, whereas
they have been appropriated as vernaculars in the settlement colonies of the Americas, Australia,
and New Zealand. The extent to which the indigenous languages have been endangered or lost in
the settlement colonies is covariant with the extent to which the indigenous populations have been
exterminated (as in the Caribbean) or the extent to which the indigenous socio-economic system has
been driven out and the Natives must function in the new socio-economic world order in order to
survive. Likewise, the extent to which those who have shifted to the new vernaculars have
influenced its structures depends largely on how early during the colonial rule (which is still in
20
In the case of sub-Saharan Africa, Côte d’Ivoire, Gabon, and Mozambique are, to my knowledge, among
the rare former exploitation colonies where large numbers of children have been reported to acquire French
or Portuguese as their vernaculars, at least in the urban environment.

process in the settlement colonies) the relevant population had to function in the new socio-
economic world order and use it for intra-group communication. Creoles are the products of
situations in which language shift occurred early and segregation was institutionalized soon
afterwards, favoring significant influence of the languages previously spoken by the African slaves,
especially the Bozals of the late plantation societies, on the structures of their new vernaculars.
Creoles have emerged as distinct vernaculars with a high degree of vitality because the African
slaves had to function quite early in the new socio-economic world order of settlement colonies.
The early imposition of socio-economic segregation created an ecology in which the slaves
socialized primarily among themselves and their colonial vernaculars could diverge structurally.
The particular ways in which their populations grew, more by importation than by birth in the 18th
and early 19th centuries, with more and more Bozal slaves acquiring the vernaculars from seasoned
slaves who had xenolectal features, paved the way for substrate elements to influence the
divergence patterns of their new vernaculars.21
At least in North America and Australia, the Natives were generally left on the margins of the
new socio-economic world orders. Trade with them was carried in the indigenous languages,
leading to the emergence of pidgin varieties such as Chinook Jargon, Mobilian, and Delaware
Pidgin (Silverstein 1996). Although the later shift of Native Americans to English, in the 19th
century, produced some sort of Native American English Pidgin (Mithun 1992), the indigenous
languages have exerted even less influence on the structures of North American Englishes than
continental European languages have. Gradual absorption by the dominant populations, marked by
the absence of Native American neighborhoods in American cities (unlike the tradition of, e.g. Irish,
Italian, German, and Black neighborhoods), led to the rapid dissolution of such transitional ethnic
varieties, in the same way that Italian and German Englishes, for instance, have vanished.
Overall, we learn that vernaculars compete with vernaculars and lingua francas with lingua
francas. What has prevented indigenous languages in Asia and Africa from being displaced by
European colonial languages is in part the fact that in exploitation colonies the latter have
functioned primarily as lingua francas, and as vernaculars only within a negligible proportion of the
elite. Interestingly, the European exploitation colonization also produced or promoted the expansion
of (new) lingua francas “lexified” by indigenous languages, e.g. Fanakalo (in South Africa), Hausa
(in Nigeria), Kikongo-Kituba, Lingala, and Sango (in the central African region), Swahili (in
eastern Africa), Town Bemba (in Zambia), and Wolof (in Senegal). Most of these function today as
urban vernaculars.
At least during the colonial period, there was no competition between the indigenous and the
European lingua francas. Quite a few lessons can also be learned in this case: 1) The European
languages have been transmitted through the school system and have been associated with the
intellectual elite. Their practice has been contained in a particular social class, thus lending more
theoretical significance to the notion of ‘population structure’. This form of segregation has
protected the indigenous lingua francas from their expansion. 2) The post-independence economic
stagnation or demise experienced by most of the African countries has halted the expansion of
European lingua francas, showing clearly that investment in a particular language can be assessed in
economics terms of costs and benefits, although the degradation of the school systems takes part of
21
The combination of contact and segregation also accounts, with a dosage of their own socio-historical
“ecological” peculiarities, for the emergence of mixed language varieties such as Michif (from French and
Creek) and Copper Island Aleut (from Russian and Aleut). See, e.g. Thomason & Kaufman (1988),
Golovko & Vakhtin (1990) and Bakker (1997) for informative discussions of their developments.
Geographical and social isolation is a critical ecological factor bearing on the development of their
structural peculiarities.

the blame in these particular cases.22 3) To be sure, the degradation of the school systems can be
blamed for the progressive indigenization of the lingua francas. This is a process that probably
would have been more rapid if the languages had vernacularized and their transmission had
depended more on naturalistic acquisition, through socialization and face-to-face interactions, like
the transmission of the indigenous lingua francas and urban vernaculars.
These considerations prompt us to reexamine the past and think again over how Vulgar Latin
spread, at a time when only a few privileged children could attend school in Europe. They also tell
us why Irish English was late in forming, although English was introduced to Ireland, as a trade
language as early as the 9th century. It is thanks to the potato plantations of the 17th century, under
Oliver Cromwell, that Ireland would be colonized on the settlement model and English would
spread informally, through naturalistic acquisition by the migrant workers. Indeed, the potato
plantations gave to English an economic value that the school system had not succeeded in doing,
not any more than it had done to Latin. This grassroots valorization spread English among the Irish
faster than the school system had within the masses of the population. The trend resulted, on the one
hand, in the indigenization and vernacularization of English and, on the other, in the gradual
extinction of Irish. More or less the same process can be assumed of Vulgar Latin as it (rather than
Classical Latin of the Intelligentsia) was valorized as the language of the new, Romanized socio-
economic system, a language that would enable the subjects of the former Roman Empire to earn a
better living.
Since European colonial languages have been maintained as official languages of their former
exploitation colonies and as lingua francas of the Intelligentsia, we can perhaps surmise the
following: Whether or not they will vernacularize and spread through naturalistic transmission
among the masses of the population depends largely on whether or not they are considered
important for regular, “blue-collar” jobs, which do not require high technical skills. (These are,
incidentally, the only jobs accessible to the vast majorities of their populations.) So far, it is the
indigenous lingua francas which are associated with such jobs. With perhaps the exception of South
Africa, the economies of Sub-Saharan Africa are far from expanding, let alone involving the rural
populations, which are still the majorities. The European colonial languages have undoubtedly
indigenized, but they have hardly vernacularized. If anything, as argued in Mufwene (2004, 2005a),
it is the indigenous urban vernaculars and regional lingua francas that are threats to the ancestral
ethnic languages, as much as they all bear essentially lexical influence from the European
languages.
5. Conclusions
Little needs to be added here that has not been said already in Part 4. If creoles have really
developed in their own unusual or abnormal way, their different structures suggest that the
ecologies of their emergence are far from being identical from one to the other. On the other hand,
if we assume a uniformitarian position and acknowledge that they have emerged by the same
restructuring processes that have often resulted in language diversification, then it is normal that
differing ecologies will produce new mutually divergent language varieties, even if exactly the
same languages are involved in the contacts. Local dynamics of competition and selection will
favor different variants even from what may appear to be more or less the same feature pools. Thus,
creoles are not genetically unusual, nor abnormal, nor less natural. Instead, they are a precious
22
This utilitarian explanation can explain, conversely, why the spread of English as a vernacular has
progressed in multilingual places like Singapore: its association with the polity’s economic progress,
favored by the small size of its population, has made English a more useful language. However, much
more remains to be learned as Cantonese seems to have slowed down a similar spread of English in Hong
Kong. Despite some general similarities, there are always local ecological specificities that must be
factored in the language evolution equation.

opportunity for linguists to realize the extent to which language contact, subsuming also dialect
contact but really based on idiolect contact, has been a critical catalyst in language change and
speciation.
This article also shows that language diversification has typically proceeded hand in hand with
language shift. Without invalidating the usefulness of the comparative method as a tool for
determining the extent to which languages share forms and structures, this article shows that genetic
creolistics can help us improve the practice of genetic linguistics and broaden usefully the scope of
its concerns. This essay is not just about the diversification and structural relatedness non-creole
languages, it is about all cases of language speciation, including those where creoles have emerged.
To the extent that contact, situated at the idiolectal level, is acknowledged as a critical ecological
factor in the actuation of change, the distinction between externally and internally-motivated change
becomes simply sociological, and the distinction between changes induced by contact and those
independent of contact becomes misguided. Moreover, like evolutionary biology, genetic linguistics
(which could also be called evolutionary linguistics) has everything to gain from being interested in
issues of language vitality, which can then deal with the demise of languages by structural erosion
or by language shift and with the emergence of new varieties. In all these cases of language
evolution, the action of competition and selection among competing variants and/or systems is
evident.
References
Aikhenwald, Alexandra Y. & Dixon, Robert M.W. (eds.). 2001. Areal diffusion and genetic
inheritance: Case studies in language change. Oxford: Oxford University Press.
Alleyne, Mervyn C. 1980. Comparative Afro-American. Ann Arbor: Karoma.
Andersen, Roger. 1990. The Papiamentu tense-aspect system, with special attention to discourse.
Pidgin and creole tense-mood-aspect systems ed. by John V. Singler, 59-96. Amsterdam:
John Benjamins.
Bailey, Charles-James N. & K. Maroldt 1977. The French lineage of English. Pidgins - creoles -
languages in contact ed. by Jurgen M. Meisel, 21-53. Tübingen: Narr.
Bailey, Guy & Natalie Maynor. 1989. The divergence controversy. American Speech 64: 12-39.
Bailey, Guy & Erik Thomas. 1998. Some aspects of African-American English phonology. African-
American English: Structure, function, and use ed. by Salikoko S. Mufwene, John R.
Rickford, Guy Bailey & John Baugh, 85-109. London: Routledge.
Baker, Philip & Adrienne Bruyn (eds.). 1998. St. Kitts and the Atlantic creoles: The texts of Samuel
Augustus Matthews in perspective. London: University of Westminster Press.
Bakker, Peter. 1997. A language of our own: The genesis of Michif, the mixed Cree-French
language of the Canadian Métis. Oxford: Oxford University Press.
Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart & Winston.
Bolton, Kingsley (ed.). 2002. Hong Kong English: Autonomy and creativity. Hong Kong: Hong
Kong University Press.
Brasseur, Patrice. 1997. Créoles à base lexicale française et français marginaux d’Amérique du
Nord. Contacts de langues, contacts de cultures, créolisation ed. by Marie-Christine
Hazaël-Massieux & Didier de Robillard, 141-166. Paris: L’Harmattan.

Campbell, Lyle. 1998. Historical linguistics: An introduction. Cambridge: Massachusetts Institute

of Technology Press.
Cassidy, Frederick G. 1980. The place of Gullah. American Speech 55: 3-16.
Chaudenson, Robert. 1973. Pour une étude comparée des créoles et parlers français d'outre-mer:
Survivance et innovation. Revue de Linguistique Romane 37: 342-371.
Chaudenson, Robert. 1979. Les créoles français. Paris: Fernand Nathan.
Chaudenson, Robert. 1990. Du mauvais usage du comparatisme: le cas des études créoles. Travaux
du Cercle Linguistique d’Aix-en-Provence 8: 123-158.
Chaudenson, Robert. 1992. Des îles, des hommes, des langues: essais sur la créolisation
linguistique et culturelle. Paris: L'Harmattan.
Chaudenson, Robert. 2001. Creolization of language and culture. London: Routledge.
Chaudenson, Robert. 2003. La créolisation: théorie, applications, implications. Paris: L’Harmattan.
Chaudenson, Robert, Raymond Mougeon & Edouard Beniak. 1993. Vers une approche panlectale
de la variation du français. Aix-en-Provence : Institut d’Etudes Créoles et Francophones.
Clyne, Michael. 2003. Dynamics of language contact: English and immigrant languages.
Cambridge: Cambridge University Press.
Corne, Chris. 1999. From French to Creole: The development of new vernaculars in the French
colonial world. London: University of Westminster Press.
Cowen, Noel. 2001. Global history: A short overview. Cambridge, UK: Polity Press.
DeCamp, David. 1971. Toward a generative analysis of a post-creole speech continuum.
Pidginization and creolization of language ed. by Dell Hymes, 349-70. Cambridge:
Cambridge University Press.
DeGraff, Michel. 1999. Creolization, language change, and language acquisition: A prolegomenon.
Language creation and language change: Creolization, diachrony, and development, ed. by
Michel DeGraff, 1-46. Cambridge, MA: Massachusetts Institute of Technology Press.
DeGraff, Michel. 2001. On the origin of creoles: A Cartesian critique of neo-Darwinian linguistics.
Linguistic Typology 5: 213-310.
DeGraff, Michel. 2003. Against creole exceptionalism. Discussion note. Language 79: 391-410.
De Landa, Manuel. 2000. A Thousand Years of Nonlinear History, Zone Books, Swerve Editions.
Faine, Jules. 1937. Philologie créole : études historiques et étymologiques sur la langue créole
d’Haïti. Port-au-Prince : Imprimerie de l’Etat.
Filppula, Markku, Juhani Klemola & Heli Pitkänen (eds.). 2002. The Celtic Roots of English,
Joensuu.
Forston, Benjamin W. 2004. Indo-European language and culture: An introduction. Malden, MA:
Blackwell.
Golovko, Eugeni V. & Nikolai Vakhtin. 1990. Aleut in contact: The CIA enigma. Acta Linguistica
Hafniensia 22: 97-125.
Goodman, Morris. 1964. A comparative study of creole French dialects. The Hague: Mouton de
Gruyter.

Goodman, Morris. 1993. African substratum: Some cautionary words. Africanisms in Afro-
American language varieties ed. by Salikoko S. Mufwene, 64-73. Athens: University of
Georgia Press.
Hagège, Claude. 1993. The language builder: An essay on the human signature in linguistic
morphogenesis. Amsterdam: John Benjamins.
Hall, Robert A., Jr. 1950. The genetic relationships of Haitian Creole. Ricerche Linguistiche 1: 194-
203.
Hall, Robert A., Jr. 1962. The life-cycle of pidgin languages. Lingua 11: 151-156.
Hall, Robert, Jr. 1966. Pidgin and creole languages. Ithaca: Cornell University Press.
Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change. Cambridge:
Hock, Hans Henrich & Brian D. Joseph. 1996. Language history, language change, and language
relationship. Berlin: Mouton de Gruyter.
Holm, John. 1989. Pidgins and creoles. Vol. 2. Reference survey. Cambridge: Cambridge
University Press.
Holm, John. 2004. Languages in contact: The partial restructuring of vernaculars. Cambrdige:
Hookoomsing, Vinesh. 2000. (Re)constructing identities: Creole and ancestral connections.
Globalization and the South-West Indian Ocean ed. by Sandra J. T. Evers & Vinesh
Hookoomsing, 149-155. Leiden, Netherlands: International Institute for Asian Studies &
Réduit, Mauritius: University of Mauritius.
Kachru, Braj. 2005. Asian Englishes: Beyond the canon. Hong Kong: Hong Kong University Press.
Kephart, Ronald F. 2000. “Broken English”: the Creole language of Carriacou. New York: Peter
Lang.
Kouwenberg, Silvia. 1994. A grammar of Berbice Dutch Creole. Berlin: Mouton De Greuter.
Kretzschmar, William A. 2002. American English: Melting pot or mixing bowl? Of dyuersitie &
chaunge of langage: Essays presented to Manfred Görlach ed. by Katja Lenz & Ruth
Möhlig, 224-39. Heidelberg, Germany: C. Winter.
Kretzschmar, William A. & Susan Tamasi. 2003. Distributional foundations for a theory of
language change. World Englishes 22: 377-401.
Labov, William. 2001. Principles of linguistic change: Social factors. Oxford, UK: Blackwell.
Labov, William & Wendell Harris. 1986. De facto segregation of black and white vernaculars.
Diversity and diachrony ed. by David Sankoff, 1-24. Amsterdam: John Benjamins.
Lafage, Suzanne. 1977. Le français en Afrique noire : Eléments de bibliographie. Abidjan :
Publications de l’Institut de Linguistique Appliquée.
Lalla, Barbara & Jean D'Costa. 1990. Language in exile: Three hundred years of Jamaican Creole.
Tuscaloosa, AL: University of Alabama Press.
Landa, Manuel De. 2000. A thousand years of nonlinear history. New York: Swerve Editions.
Lass, Roger. 1997. Historical linguistics and language change. Cambridge: Cambridge University
Press.

Le Page, Robert .B. & Andrée Tabouret-Keller. 1985. Acts of Identity: Creole-based approaches to
language and identity. Cambridge: Cambridge University Press.
Lightfoot, David W. 2002. Myths and the prehistory of grammar. Journal of Linguistics 38: 113-
136.
Manessy, Gabriel. 1994. Créoles, pidgins, variétés véhiculaires: procès et genèse. Paris: Centre
National de la Recherche Scientifique.
Manessy, Gabriel & Paul Wald. 1984. Le français en Afrique noire, tel qu’on le parle, tel qu'on le
dit. Paris: L'Harmattan.
McMahon, April M. 1994. Understanding language change. Cambridge: Cambridge University
Press.
McWhorter, John H. 1998. Identifying the creole prototype: Vindicating a typological class.
Language 74: 788-818.
McWhorter, John H. 2000. The missing Spanish creoles: Recovering the birth of plantation contact
languages. Berkeley: University of California Press.
McWhorter, John H. 2001. The power of Babel: a natural history of language. New York: Times
Books.
McWhorter, John H. 2005. Defining Creole. Oxford: Oxford University Press.
Meillet, Antoine. 1982 [1900]. Note sure une difficulté générale de la grammaire comparée.
Reprinted in Linguistique historique et linguistique générale, 36-43. Genève : Slaktine &
Paris : Champion.
Meillet, Antoine. 1921. Linguistique historique et linguistique générale. Vol. 1. Paris: Champion.
Meillet, Antoine. 1929. Le développement des langues. In Continu et discontinu, 119ff. Paris:
Bloud & Gay. Reprinted in Meillet 1951: 71-83.
Meillet, Antoine. 1951. Linguistique historique et linguistique générale. Vol. 2. Paris: Klincksieck.
Mithun, Marianne. 1992. The substratum in grammar and discourse. Language contact: Theoretical
and empirical studies ed. by Ernst Håkon Jahr, 103-115. Berlin: Mouton de Gruyter.
Mufwene, Salikoko S. 1988. Why study pidgins and creoles? Column. Journal of Pidgin and
Creole Languages, 3: 265-76.
Mufwene, Salikoko S. 1991. Is Gullah decreolizing? A comparison of a speech sample of the 1930's
with a speech sample of the 1980’s. The emergence of Black English ed. by Guy Bailey,
Patricia Cukor-Avila & Natalie Maynor, 213-30. Amsterdam: John Benjamins.
Mufwene, Salikoko S. 1993. African substratum: Possibility and evidence. Africanisms in Afro-
American language varieties ed. by Salikoko S. Mufwene, 192-208. Athens: University of
Georgia Press.
Mufwene, Salikoko S. 1994. On decreolization: The case of Gullah. Language and the social
construction of identity in creole situations ed. by Marcyliena Morgan, 63-99. Los Angeles:
Center for Afro-American Studies.
Mufwene, Salikoko S. 1996. The Founder Principle in creole genesis. Diachronica 13: 83-134.
Mufwene, Salikoko S. 1997a. Jargons, pidgins, creoles, and koinés: What are they? The structure
and status of pidgins and creoles ed. by Arthur K. Spears & Donald Winford, 35-70.
Amsterdam: John Benjamins.

Mufwene, Salikoko S. 1997b. The legitimate and illegitimate offspring of English. World Englishes
2000 ed. by Michael L. Forman & Larry Smith, 182-203. Honolulu: The University of
Hawaii Press.
Mufwene, Salikoko S. 2000a. Creolization is a social, not a structural, process. Degrees of
restructuring in creole languages ed. by Ingrid Neumann-Holzschuh & Edgar Schneider,
65-84. Amsterdam: John Benjamins.
Mufwene, Salikoko S. 2000b. Population contacts and the evolution of English. The European
English Messenger 9: 9-15.
Mufwene, Salikoko S. 2001. The ecology of language evolution. Cambridge: Cambridge University
Press.
Mufwene, Salikoko S. 2003a. Genetic linguistics and genetic creolistics. Short note. Journal of
Pidgin and Creole Languages 18: 273-288.
Mufwene, Salikoko S. 2003b. Competition and selection in language evolution. Selection 3.45-56.
Mufwene, Salikoko S. 2003c. Colonization, globalization, and the future of languages in the
twenty-first century. MOST Journal of Multicultural Societies 4,
http://www.unesco.org/most/vl4n2mufwene.pdf
Mufwene, Salikoko S. 2004. Multilingualism in linguistic history: Creolization and indigenization.
Handbook of bilingualism ed. by Tej Bhatia and William Ritchie, 460-488. Malden, MA:
Blackwell.
Mufwene, Salikoko S. 2005a. Créoles, écologie sociale, évolution linguistique. Paris: L’Harmattan.
Mufwene, Salikoko S. 2005b. Language evolution: The population genetics way. Gene, Sprachen,
und ihre Evolution ed. by Guenther Hauska, 30-52. Universitaetsverlag Regensburg.
Mufwene, Salikoko S. to appear. Race, racialism, and the study of language evolution in America.
LAVIS III – Language Variety in the South: Historical and Contemporary Perspectives ed.
by Michael Picone & Katherine Davis. Tuscaloosa: University of Alabama Press.
Naro, Anthony Julius & Maria Marta Pereira Scherre. 1993. Sobre as origins do Português popular
do Brasil. Revista de Documentação de Estudos em Lingüística Teórica e Aplicada
(DELTA) 9: 439-545.
Nettle, Daniel. 1999. Using social impact theory to simulate language change. Lingua 108: 95-117.
Pargman, Sheri. 2002. Internal and external factors in language change. PhD dissertation,
University of Chicago.
Pollard, Velma. 2000. Dread Talk: The language of Rastafari. Mona, Jamaica: The University of
the West Indies Press.
Polomé, Edgar. 1983. Creolization and language change. The social context of creolization ed. by
Ellen Woolford and William Washabaugh, 126-36. Ann Arbor: Karoma.
Posner, Rebecca. 1985. Creolization as typological change: Some examples from Romance syntax.
Diachronica 2: 167-88.
Prescod, Paula. 2004. Une description grammaticale du syntagme nominal dans le créole
anglophone de St-Vincent-et-les-Grenadines. Thèse de doctorat, Université Paris III −
Sorbonne Nouvelle.

Rickford, John R. & Jerome S. Handler. 1994. Textual Evidence of the Nature of Early Barbadian
Speech. Journal of Pidgin and Creole Languages 9: 221-255.
Schlieben-Lange, Brigitte. 1977. L’origine des langues romanes: un cas de créolisation? In Langues
en contact - pidgins - creoles - Languages in contact, ed. by Jürgen Meisel, 81-101.
Tübingen: Gunter Narr.
Schneider, Edgar W. 1990. The cline of creoleness in English-oriented creoles and semi-creoles of
the Caribbean. English World-Wide 11: 79-113.
Schuchardt, Hugo. 1914. Die Sprache der Saramakkaneger in Surinam. Amsterdam: Johannes
Muller.
Schwartz, Stuart B. 1985. Sugar plantations in the formation of Brazilian society: Bahia, 1550-
1835. Cambridge: Cambridge University Press.
Schwegler, Armin. 2006. Bozal Spanish: Captivating new evidence from a contemporary source
(Afro-Cuban “Palo Monte”). Studies in Contact Linguistics: Essays in Honor of Glenn G.
Gilbert, ed. by Janet Fuller and Linda L. Thornburg,71-101. New York: Peter Lang.
Silverstein, Michael. 1996. Dynamics of recent linguistic contact. Handbook of North American
Indians vol. 17, Languages, ed. by Ives Goddard, 117-136. Washington, DC: Smithsonian
Institution Press.
Thomason, Sarah G. 2001. Language contact: An introduction. Washington, DC: Georgetown
University Press.
Thomason, Sarah G. (2002). Creoles and genetic relationship. Column. Journal of Pidgin and
Creole Languages 17: 101-109.
Thomason, Sarah G. 2003. A response to Mufwene’s response [Mufwene 2003a]. Journal of Pidgin
and Creole Languages 18: 289-298.
Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization, and genetic
linguistics. Berkeley: University of California Press.
Trask, R.L. (1996). Historical linguistics. Londres: Arnold.
Trubetzkoy, Nikolai S. 1939. Gedanken über das Indogermanenproblem. Acta Linguistica 1: 81-89.
Trudgill, Peter P. 2004. New-dialect formation: The inevitability of colonial Englishes. Oxford:
Oxford University Press.
Weinreich, Uriel, William Labov & Marvin I. Herzog. 1968. Empirical foundations for a theory of
language change. Directions for historical linguistics ed. by Winfred P. Lehman & Yakov
Malkiel, 97-195. Austin: University of Texas Press.
Winford, Donald. 1997. Re-examining Caribbean English creole continua. World Englishes 16:
233-279.


Population Movements and Contacts

Uploaded by

Copyright:

Available Formats

Population Movements and Contacts

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Population Movements and Contacts

Uploaded by

Copyright:

Available Formats

POPULATION MOVEMENTS AND CONTACTS IN LANGUAGE

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Map 1: Complementary geographic distribution of creoles and pidgins

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Map 2: European Colonial Expansion since the 15th century

Map 3: Indo-European Dispersal: The first colonial expansion

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Map 4: A Geographically-Adapted Indo-European Cladogram

Journal of Language Contact – THEMA 1 (2007)

Map 5: The Non-Rectilinear Expansion of the Indo-Europeans (same as Map 3)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Campbell, Lyle. 1998. Historical linguistics: An introduction. Cambridge: Massachusetts Institute

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

Journal of Language Contact – THEMA 1 (2007)

You might also like