See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343099508

The Middle English Creolization Hypothesis: Persistence, Implications, and

Language Ideology

Article in Studia Anglica Posnaniensia · March 2019

DOI: 10.2478/stap-2019-0006

CITATIONS READS

0 2,310

1 author:

David O'Neil
University of Southern Indiana
8 PUBLICATIONS 52 CITATIONS

SEE PROFILE

All content following this page was uploaded by David O'Neil on 21 July 2020.

The user has requested enhancement of the downloaded file.

 Studia Anglica Posnaniensia 54 (2019): 113–132
doi: 10.2478/stap-2019-0006

THE MIDDLE ENGLISH CREOLIZATION HYPOTHESIS:

PERSISTENCE, IMPLICATIONS, AND LANGUAGE IDEOLOGY

DAVID O’NEIL1

ABSTRACT

Bailey and Maroldt (1977) and Domingue (1977) were the first to argue that language contact
during the Middle Ages between Old English and both Old Norse and Norman French resulted in
linguistic creolization. This theory, known as the Middle English creolization hypothesis, implies
that Middle English, and perhaps Modern English as well, should be classified as a creole. Though
frequently discredited on historic, linguistic, and terminological grounds, the creolization
hypothesis has attracted interest for longer than might be expected. This paper argues that the
persistence of the hypothesis may be ideologically motivated. The first section examines
connotations of the term “creole” and applies these connotations to an analysis of the initial
presentations of the creolization hypothesis. The second and third section of the paper review and
analyze the forty-year history of the debate, focusing separately on arguments for creolization (and
koinezation) between Anglo-Norman French and Old Norse, respectively. The fourth and final
section examines challenges presented by the concept of creole exceptionalism to common attitudes
about language equality and the theory of Universal Grammar. It is argued that these issues attract
greater interest when contextualized within a discussion of a “major” world language such as
English than when creolization is understood as an atypical process restricted to “peripheral”
languages such as Haitian Creole. This paper also references relevant political issues such as the
current controversy among medievalists about the field’s historic lack of inclusivity.

Keywords: Middle English; creolization hypothesis; creole; koine; history of English linguistics;
language ideology.

1. Introduction

The term “creole” has been used by linguists and other language scholars for a
long time, yet consistent definition remains elusive. According to Görlach,
specialists have frequently redefined the term “to make it satisfy the specific

1 David O'Neil, University of Southern Indiana, 8600 University Boulevard, 47712 Evansville,
IN, United States, [email protected]
114 D. O’Neil

needs of their arguments” (1986: 330), and McWhorter argues that “there
remains a gaping lack of consensus in the subfield as to what a creole even is”
(2005: 37). As might be expected, confusion is even more likely among non-
specialists. Watts recounts a debate from an internet-based language forum in
which practically everyone had his or her own definition. Some participants
believed that creoles were by definition the result of European colonialism (for
some, only French colonialism), and others argued that a creole was just a
simplified language, a “bastard” that resulted from extended language contact
regardless of geographical, cultural, or social context. Only one commenter
offered the classical definition that a creole was “a pidgin language which has
become a mother tongue” (2011: 86–87). Clearly, a definition for the term has
not been easy to agree on. Even among specialists, determining which languages
should be classified as creoles has often entailed significant and lasting
controversy. One example of such controversy is the forty-year debate over the
Middle English creolization hypothesis.
The presence of the Middle English creolization hypothesis in the scholarly
literature has been durable, yet paradoxically the hypothesis itself is not popular
among language scholars. From its first articulation in the late seventies, it has
faced strong and repeated criticism on terminological, historic, and linguistic
grounds – as seen, for example, in responses by Poussa (1982), Görlach (1986),
Thomason & Kaufman (1988), Dalton-Puffer (1995), Allen (1997), Danchev
(1997), Hogg (1997), McWhorter (2005), Watts (2011), and Trotter (2017). The
fact that interest has continued in the face of such opposition suggests there is
more involved than the linguistic classification of a single language. In this paper,
I argue that the mere idea of classifying English as a creole has ideological
resonance that provokes our conception of English as a language, as well as our
general understanding of language and language change. Most obviously, the
hypothesis entails classifying a “major” European language alongside “minor”
Caribbean languages such as Haitian Creole and Jamaican Patois, and for many
scholars this association may be either intriguing or uncomfortable. Less
obviously, creolization presents challenges to general assumptions about
linguistic inheritance and linguistic complexity – challenges which are harder to
ignore when contextualized within the discussion of a language such as English,
raising the stakes for the debate in ways that might not be expected. In Section 2
I discuss why creolization debates are so often ideologically charged, focusing
on claims made by the first proponents of the hypothesis. In Sections 3 and 4
I present overviews of linguistic and historical arguments surrounding the
hypothesis, noting ramifications of the discussion that extend beyond the narrow
limits of the field. Finally, in Section 5 I discuss ideological implications of
creolization and creole exceptionalism as they relate to our conception of the
general human language capacity.
 The Middle English creolization hypothesis 115

2. Creolization and language ideology

Given its connotative associations, conversations surrounding creoles and the

process of creolization are often ideologically charged. DeGraff (2003), for
example, argues that the term has been incurably tainted by its origins in
nineteenth-century colonialism – and he has a point. Consider the frank racism in
what passed as a scholarly discussion of creoles in 1889:

Creole languages result from the adaptation of a language, especially some Indo-
European language, to the (so to speak) phonetic and grammatical genius of a race
that is linguistically inferior. The resulting language is composite, truly mixed in its
vocabulary, but its grammar remains essentially Indo-European, albeit extremely
simplified. (after DeGraff 2003: 393)

Frank comments such as this one about linguistic inferiority and the “grammatical
genius of a race” are, for the most part, long gone from the field, but this
disciplinary history has cast a long shadow over modern-day conversations about
creolization. Some scholars are understandably hesitant to ascribe special
characteristics to languages spoken by colonized peoples, who have a long history
of being categorized as “other.” This history may partly explain why discussions
of creoles and creole theory may attract greater interest than might be expected
from their otherwise narrow academic content, especially when the debate is over
a language such as English that has tended to fill the lexifier role in accounts of
creolization. Merely to contemplate the idea of English as a creole has the power
to challenge stereotypes, subvert the problematic early history of creole studies,
and foreground questions that might otherwise have been neglected.
The hypothesis that Middle English was a creole originated in the work of
Bailey and Maroldt (1977) and Domingue (1977), who independently argued
that the linguistic change evident in the transition from Old to Middle English
should be attributed to language contact and creolization. In the first appearance
of the hypothesis, Bailey and Maroldt argued for two accounts of this process,
which were based on grammatical influence and the frequency of loan words in
English. The first account was that Middle English was a creole that resulted
from the mixing of Old English and Norman French after the mid-eleventh-
century invasion of England under William the Conqueror. In short, French
acted as superstrate, or lexifier, to an Old English substrate, and the language
contact continued as long as speakers of Norman French filled positions of
power in England (until about the thirteenth or fourteenth century). Bailey and
Maroldt also claimed that, even prior to the creolization of Middle English, Old
English had already undergone creolization during its ninth- and tenth-century
contact with Old Norse. Together, these claims imply English was creolizing
for around 500 years, which is a somewhat surprising assertion considering the
116 D. O’Neil

conventional view that creolization is a rapid process that takes place within a
few generations.
But Bailey and Maroldt rely on a broader definition of creoles. As they put it,
creoles are “the result of mixing which is substantial enough to result in a new
system, a system that is separate from its antecedent parent systems” (1977: 21).
This definition is criticized by Görlach (1986) on the grounds that it would allow
almost any language to be classified as a creole, but it’s doubtful Bailey and
Maroldt would mind this criticism given their claim that all languages should be
represented, “like humans,” as having two parents rather than one (1977: 22).
With this claim, the authors reject the common view, represented in standard
dendrograms, that languages descend from a single ancestor. They explicitly
describe creolization not as a special linguistic process, but as a commonplace
mechanism by which most if not all languages change over time. More than just
our conception of English, Bailey and Maroldt were challenging our
understanding of how, when, and where creolization occurs as a general process,
and the current social, cultural, and political status of English made Middle
English an ideal battleground for arguing that creoles are non-exceptional
languages. Thus, even in its earliest expression, the creolization hypothesis was
accompanied by a frankly ulterior motive: the desire to show creolization as a
common, or even universal process of language change that all languages are
subject to. In this work, the term “creole” seems to have been adopted, not for
classificatory precision, but rather to highjack its rhetorical force.
Domingue independently proposed a creolization hypothesis the same year as
Bailey and Maroldt, arguing that Middle English was not a “modernization” of
Old English, but a “linguistic innovation” that resulted from contact with several
languages (1977: 89). As support, she points to the very high frequency of French
words in the Middle English lexicon and also to the smaller number of words of
Scandinavian origin, which nevertheless account for a significant contribution to
the lexical core, where loan words are less likely to be found (e.g., they, them,
their, are, till, though, etc.). Domingue also finds evidence for creolization in
Middle English’s impoverished inflectional system, which she argues could have
resulted either from the “convergence of several systems” or a “general trend
toward simplification following universal rules” (1977: 92–93). In either case,
she explains that contact with French and Old Norse drove the shift in English
language syntax from relatively synthetic (Old English) to analytic (Middle
English).
There are a couple of key differences between Domingue’s and Bailey and
Maroldt’s understanding of creoles. One difference is that Domingue takes the
orthodox position that a creole must develop from a pidgin (though no such
pidgin exists in the historical record in the case of English). A more important
difference, however, is Domingue’s reluctance to make any definitive claims. At
 The Middle English creolization hypothesis 117

the outset, she goes no further than to state that Middle English is “very much
alike a creole” (1977: 89) [emphasis mine], but she concludes with a serious
consideration of the implications entailed by either strong position in the debate.
On one hand, denying Middle English creole status could “be very helpful if we
want to define the term creole more narrowly than we have done up to now”
(1977: 97). On the other hand, if the claim of creolization is accepted for Middle
English, then the historical processes of pidginization and creolization may be
more widespread than usually considered. Thus, the designation of Middle
English as a creole could impact our understanding of historical processes of how
language families are constituted:

It can then be argued that pidginization and subsequent creolization may be much
more common than usually held. They may be widely spread processes in the
history of languages… Proto-Germanic could be considered a creole language, as
could Proto-Armenian and Proto-Albanian. Processes of pidginization and
creolization can explain better than borrowing or areal influence the rationale
behind the wave model for a classification of Indo-European. It might also well be
that genetic classifications based on lexico-phonological correspondences are not
only too rigid, but misleading, hiding facts of pidginization/creolization, because
they do not take syntactic features into consideration. Such speculations strip
pidgins and creoles of their status of ‘special’ languages and regard them as normal,
though not necessary, steps in the formation of languages. (1977: 97)

In this discussion, Domingue speculates about the broader implications of the

creolization debate, which are argued to extend well beyond English itself. Like
Bailey and Maroldt, she sees the debate as a battlefield for larger issues in the
field of historical linguistics. If English can be classified as a creole, she suggests,
then maybe creoles are not exceptional.
Unlike Bailey and Maroldt, Domingue hesitates to defend the strongest form of
the hypothesis, but she also places the burden of proof on those who would deny
Middle English the status of creole. Is there “a specifically creole syntax which ME
does not possess?” she asks (1977: 97). In the end, Domingue remains agnostic and
suggests reserving the term “creole” for those languages in the West Indies and
Indian Ocean whose speakers use the term for themselves. As an alternative, she
proposes the term “hybrid” for languages that exhibit significant mixing – a
solution that may avoid some of the terminological problems in creole theory,
though at the cost of ignoring the question of whether pidginization and creolization
really are distinct processes that explain the origins of certain languages. After all,
if the term “creole” is divorced from the classical definition of “nativized pidgin”,
then not only does the term lose meaning, but an important concept may be lost.
Markey offers a pointed observation about such attempts to redefine the term:
“To label creóles ‘contact languages’ is vacuous: all languages are contact
languages. To call all languages creóles is equally fatuous…” (1982: 170). Of the
118 D. O’Neil

initial authors of the creolization hypothesis, Domingue is guilty of the first extreme
(simplifying the term “creole” to “contact language”), and Bailey and Maroldt are
guilty of the second (labeling all languages as creoles).
While Domingue hoped to circumvent the rhetorical challenges of the term
“creole” by proposing a new term, for Bailey and Maroldt these challenges are
rather the point. They openly object to standard views of linguistic inheritance
and frankly contend that the standard tree model of linguistic ancestry should be
rewritten to show lineage from multiple parents. To say the least, such a drastic
revision of how language families operate would disrupt theories based on the
standard model. Whether explicitly or implicitly, linguists tend to think of creoles
and non-creoles as being derived from two completely distinct models of
parentage. Bailey and Maroldt reject this distinction, but DeGraff (2003) makes
an even more passionate case for the insidiousness of views that interpret
creolization as an atypical process of linguistic inheritance. Since creoles and
pidgins are segregated into their own “genetic” group marked by a lack of
grammatical structure, he sees creole theory as a continuation of nineteenth-
century colonialist attitudes toward “inferior” races:

[P]idginization is an all-powerful structural simplification process that is claimed

to, among other things, obliterate virtually all morphology and to lead to a pre-
Creole early contact language—a ‘macaronic’ jargon or ‘early pidgin’—whose
extraordinary lack of structure makes it unlike any full-fledged human language
and more like some protolanguage that may have been spoken by our (pre)-hominid
ancestors. (2003: 398)

The main force of DeGraff’s argument is that creole exceptionalism encodes

historical prejudices, masking imperialist attitudes from a century ago under a
veneer of technical argumentation. This point may be pushed even further. If
pidgins are argued to have no direct ancestors in their genealogy, then both
pidgins and creoles would have no place in a linguistic model that presents
languages as related by common descent, casting creoles in an extreme position
of genetic alterity.
So what happens if English, a language deeply associated with what Milroy
calls the “ideology of the standard language” (2001a: 530), should be determined
to be a creole? As I have argued, it may be that the creolization hypothesis for
Middle English has been persistent in scholarly debate precisely because of its
rhetorical power to provoke such challenges. English is a “major” member of the
Indo-European language family, and its place on the world stage is currently so
dominant that it has commonly been described as imperialistic (for linguistic
imperialism, see Phillipson 1992; for the current role of English in international
publishing, see O’Neil 2018). For this reason, issues regarding the classification
of English are harder for the scholarly community to ignore than issues regarding
 The Middle English creolization hypothesis 119

“minor” colonial languages. This would be true even if the theoretical challenges
for English were no different than they were for languages that are considered
creoles uncontroversially. Even in its earliest framing, therefore, the creolization
hypothesis was presented within an explicit ideological context. Bailey and
Maroldt, in particular, were making a cross-linguistic, universal claim about
linguistic inheritance. I argue that they picked English as the battleground in order
to make other scholars pay attention.
No doubt, the strategy worked, but neither Bailey and Maroldt nor Domingue
find much support in the subsequent literature, especially with regard to the
French side of the hypothesis. In fact, Plag criticizes a fellow scholar for wasting
time bothering to refute the hypothesis, insisting that it “has long been dismissed
as thoroughly inadequate” (1998: 393). But I disagree that the hypothesis is
unworthy of our attention. Despite its weak and perhaps intentionally iconoclastic
presentation in some of the literature, the response has spurred interesting and
productive work on the development of Middle English. Furthermore, it is
sensible for scholars to be cautious when addressing a question with ideological
ramifications. As a parallel case, Markey points out that the most strident
opposition to defining Afrikaans as a creole has been associated with
eurocentrism and white supremacy (in this context, creolization was described as
“linguistic miscegenation,” 1982: 170). The contact situation for English in the
Middle Ages is not the same as colonial South Africa, but careful review of the
evidence is merited when there may be even the faintest suspicion of problematic
motives. And medievalists have reason to be especially sensitive right now. The
field is currently embroiled in crisis about its lack of inclusivity, as illustrated by
recent headlines from The New York Times and The Washington Post (just two of
many examples). The headlines read, respectively, “Medieval Scholars Joust with
White Nationalists. And One Another” (2019, May 5) and “It’s All White People:
Allegations of White Supremacy are Tearing Apart a Prestigious Medieval
Studies Group” (2019, September 15). The first article addresses the general
controversies currently shaking up the field; the second discusses a particular
controversy centering on the connotative resonance of the term “Anglo-Saxon,”
which some critics argue should be discarded on the grounds that it has become
a “code for whiteness”. Again, the fact that “creole” has related ideological
connotations (in a sense, nearly the converse connotations) helps explain the
surprising degree of interest in the Middle English creolization hypothesis.

3. Anglo-Norman French

The strongest linguistic evidence against Middle English creolization is found in

the arguments related to Anglo-Norman French (as opposed to Old Norse). One
of the earliest critics, Poussa objects to the hypothesis’s overreliance on evidence
120 D. O’Neil

of lexical borrowing, maintaining that even large influxes of loanwords and

translation calques are no proof of creolization in the usual sense unless there is
also “evidence of more fundamental systemic change" (1982: 70). McWhorter
makes much the same point, adding that even morphological borrowing is not
conclusive (2005: 267). Obviously, the lack of a common set of criteria can make
these arguments difficult to evaluate. Dalton-Puffer insists that the presence of
similar features among languages in a contact situation is not enough for reliable
classification “unless, of course, we are not too worried about being imprecise in
our terminology” (1995: 48). In this, she shares Domingue’s concern for
establishing stricter definitions in creole studies. If all languages that result from
contact are considered creoles, then perhaps the term is too diluted to serve any
purpose outside its rhetorical function.
To illustrate the importance of caution in these arguments, Dalton-Puffer’s
employs a compelling argumentum ad absurdum, in which she shows that
Middle English exhibited many similarities to dying languages. Examples of
these features are the loss of traditional personal names, the reduction of
productivity in word formation rules, and massive one-directional lexical
borrowing. To be more specific on this last point, English absorbed massive
amounts of Romance vocabulary from Norman French, while Norman French
borrowed almost no words from English. In many of these cases, English did
not lack the word for a particular concept, but gave up pre-existing English
terminology in favor of Romance alternatives, such as in the case of merci and
compassioun for the Old English mildheortnesse (1995: 43). Again, Dalton-
Puffer marshals this evidence in support of a conclusion that is demonstrably
false (i.e., that Middle English was a dying language). The point is that
historical conclusions based on speculative or circumstantial evidence are likely
to be at odds with what actually occurred.
In addition to making this argumentum ad absurdum, Dalton-Puffer identifies
specific problems with the view that Middle English resulted from creolization
with French. For one, while a significant percentage of vocabulary in Middle
English comes from French (as many as 50 percent of the distinct lexical entries
represented in Chaucer), these borrowings were not used with nearly the same
regularity as words of Old English origin. The actual number of French tokens in
Middle English texts is quite small (in Chaucer, as low as 7 percent), because
these tokens did not come from the lexical core. In other words, there are many
lexemes in Chaucer borrowed from French, but these lexemes are infrequently
used. As Dalton-Puffer argues, since the majority of Middle English vocabulary
came from Old English, it makes little sense to consider Norman French a
superstrate language and lexifier.
Besides changes in its lexicon, another important argument for creolization in
English is the “sudden” impoverishment of inflectional morphology. However,
 The Middle English creolization hypothesis 121

Allen presents systematic manuscript evidence that the process of paradigmatic

leveling appears to have preceded language contact with French or even Old
Norse (although the contact in the latter case may have accelerated the process):
“[T]he appearance of very rapid change in EME is to a large extent illusory, as
the conservative practices of OE scribes often masked substantial changes which
had already taken place before the Norman invasion” (1997: 64–65). Specifically,
Allen demonstrates that syncretism in the nominal system was already well
advanced during the Old English period and that the inflectional loss in northern
Middle English was catalyzed by phonological changes rather than language
contact. Moreover, similar changes in southern dialects in the eleventh century
(such as loss of final ‘n’ and vowel reduction) could have led to the increasing
use of the uninflected determiner since phonological distinctions became difficult
or even impossible to perceive. At the same time, nominative-accusative
distinctions became particularly weak (though authorial understanding of the
basic inflectional system remains clear in many manuscripts, suggesting that
syncretism was not the result of new speakers being confused about
morphological categories). All these points find support in Hogg (1997), whose
only refinement of Allen’s position is to note that she underestimates the extent
of pre-contact category loss, which strengthens her main point that inflectional
loss in Middle English should not be taken as support for creolization.
The claim that morphosyntactic changes were occurring in English even
before the Norman invasion is supported by evidence outside the creolization
debate. In recent years, there has been growing interest in syntactic change as an
explanation for the metrical evolution of the medieval English alliterative
tradition (for example, in Hartman 2016; O’Neil 2017; Russom 2017). The
argument can be summarized as follows: Parallel with the case syncretism
discussed above, functors (words from closed-class, functional parts of speech
such as pronouns and prepositions) increased in frequency to compensate for the
meaning that had been expressed in the inflectional endings. Functors are
composed of unstressed syllables, so this increase influenced the distribution of
stressed and unstressed syllables in the alliterative line. Metrical analysis has
shown that these changes were already evident during the Old English period.
Hartman (2016) and Russom (2017) analyze the role of functors in the late Old
English poem The Battle of Maldon, and O’Neil (2020, forthcoming) argues that
syntactic change was appearing even earlier, showing a trend across six Old
English poems leading up to the Middle English Brut. These data show that the
oldest English alliterative poem, Cædmon’s Hymn, contained an average of about
one functor per two half-lines; mid-Old English poems like Beowulf and Andreas
contained an average of about one functor per half-line; Maldon contained an
average of about one-and-a-half functors per half-line; finally, the early Middle
English Brut contained about two functors per half-line. Applied to the
122 D. O’Neil

creolization debate, this evidence supports a model of gradual syntactic change

rather than rapid creolization.
In addition to these linguistic arguments, the creolization hypothesis has also
been disputed by means of historical reconstructions of the contact situation.
Pidgins and creoles are not generally thought of as the product of literary contact,
but of interpersonal communication in specific social situations. Yet our Middle
English texts are not, as Dalton-Puffer points out, “classified as ‘speech-based’ let
alone ‘spoken’” (1995: 38). Instead, we need to consider the extent of interpersonal
contact between speakers, and it seems likely that the contact between speakers of
English and Norman French, due to the great social distance, would not have been
the sort of contact that leads to the creation of a pidgin. This view is taken by
McWhorter (2005) and Thomason and Kaufman (1988), who argue that there was
too little contact between the French-speaking Norman invaders and the native
English speakers to have had a significant effect on the grammatical characteristics
of Middle English. To illustrate the point, Watts (2011) imagines a few language
contact scenarios that would have been likely in post-invasion England. In one, an
English-speaking peasant is called to testify before a French-speaking magistrate.
This is a poor context for language contact because the peasant would not be
expected to learn French for such a rare occurrence, while the French-speaking
magistrate might simply have hired an interpreter. In a second scenario, an English-
speaking clerk discusses financial accounts with a French-speaking lord. This is a
poor context because the social distance between them would have put the onus of
language learning on the clerk. Communication in French (even if poor French)
seems more likely in this case than the mutual adoption of a pidgin. A
counterargument could be made that extended contact between English and French
speakers even in a small number of contexts (such as trade or legal matters) could
have been enough to induce the creation of a pidgin. However, even if such a pidgin
were to exist, only very small numbers of the English-speaking population would
have had cause to use it. It seems improbable that the speech of such a group, the
members of which obviously would not have spoken with other English speakers
in the pidgin, could explain the radical transformation of the English language in
the Middle English period.
Ingham (2014) provides an explanation for the widescale transfer of French
vocabulary into Middle English that is much more compelling than the
creolization hypothesis. Countering claims that French was a language of
instruction during the middle of the Middle English period, he provides evidence
that English speakers from the educated class became bilingual in Anglo-Norman
French during middle childhood before they began formal language instruction
in Latin (which they learned via French). Bilingualism, rather than contact
between speakers of different languages, would thus have provided the basis for
French influence on English:
 The Middle English creolization hypothesis 123

Now French language influence on English is known to have been at its zenith
between 1250 and 1350 (Dekeyser 1986), by which time monolingual French
speakers were scarcely to be found, but bilingual speakers were the norm amongst
the educated classes, so this was a period when French could indeed influence
English via bilingual competence. Hence the known linguistic facts as regards
French loans in English, and the acquisitional milieu we have argued for here, are
in diachronic alignment. It is far less plausible to suppose that French could have
influenced English so strongly, had it been merely an instructed foreign language,
a status which it had in the 16th century (Kibbee 1991), when French loanword
frequency had already sharply declined. (2014: 444)

At this period, French speakers in England would have been overwhelmingly

bilingual, and it makes little sense to imagine a novel contact language arising
between monolingual and bilingual speakers of English. Accordingly, the
hypothesis that Middle English is a creole of French and Old English is rather
weak on both linguistic and historical grounds.

4. Old Norse

As compared to contact between English and French, the technical case for
creolization with Old Norse has some strengths. First, the historical setting was
much more conducive to sustained contact between Old English and Old Norse
since Danish and English farmers would have interacted on a regular basis for
two centuries in the area of northern England known as the Danelaw. Notably,
the northern dialects of Middle English from these areas show more advanced
signs of morphological leveling (discussed above) than elsewhere in England,
and inflectional loss is an essential part of any compelling argument for
creolization. If we return to Watts and his imagined scenarios for linguistic
contact, we can see that speakers in this area would be more likely to have
interacted in a setting that might necessitate a common language. Although the
image may be somewhat playful, Watts envisions an English and Scandinavian
farmer sharing a jug of alcohol and chatting amicably after a day at the cattle
market. In such scenarios, the contact would have been strictly oral rather than
literary, and Dalton-Puffer argues that it is unlikely there was “any contact
between written forms of Scandinavian and English” (1995: 48). In short,
communication between English and Danish subjects of the Danelaw was likely
to have been overwhelmingly interpersonal – a good context for creolization.
For these reasons, an early creolization process in the north (which also finds
support in Poussa 1982) is much more plausible than a later process based on
several centuries of English commoners interacting with Norman aristocrats.
However, plausibility is hardly a decisive consideration. According to Thomason
and Kaufman, the “pervasive” Norse influence was shallow and unevenly
distributed in all areas of the language aside from the lexicon, and the similarities
124 D. O’Neil

between Norse and English precluded the possibility of the former making a
significant mark on the typology of the latter (1988: 302–303). On the other hand,
McWhorter claims that an “incomplete acquisition of English” among
Scandinavians learning English as a second language could explain the early
inflectional loss characteristic of northern dialects (2005: 305), and similarities
between the languages would not eliminate this effect. This argument is presented
with even greater clarity and force in McWhorter (2008).
If this is indeed the case, many of the grammatical features of Middle English,
such as the decline of derivational prefixes and the loss of the indefinite man
(discussed in McWhorter), might be explained by features of Old Norse, even if
the context is not exactly what we would expect for creolization (though some of
the features lost in English are found in modern Scandinavian languages). The
issue here, as so often in this debate, may be terminological. Since Old English
and Old Norse are so closely related, perhaps “koiné” is a more appropriate term
for Middle English than “creole”. Watts argues that the contact between speakers
of English and either Danish or Old Norse in the Danelaw would have created
linguistic forms “similar to those resulting from processes of koïneisation in
which two mutually intelligible language varieties contribute towards a new
variety over a period of roughly three generations of speakers” (2011: 98–99)
[emphasis mine]. Although Watts (like Domingue, above) hedges his bets, the
argument is quite good. In a general discussion of historical linguistics, Hock and
Joseph identify three conditions that must be met for the establishment of a koiné:
(1) the languages in question are closely related or even mutually intelligible
dialects; (2) the languages are of about equal prestige; and (3) no outside language
is available as a link language (1996: 388). All three of these conditions are met
in the context of the ninth- and tenth-century English Danelaw. Siegel’s definition
also emphasizes the significance of mutual intelligibility:

A koine is a stabilized contact variety which results from the mixing and subsequent
levelling of features of varieties which are similar enough to be mutually
intelligible, such as regional or social dialects. This occurs in the context of
increased interaction or integration among speakers of these varieties. (2001: 175)

Evidence for mutual intelligibility can be found in Townend’s (2002) book-

length study of Anglo-Norse contact during the Viking Age, in which modern
methods are adapted to investigate the historical situation. In addition to
providing evidence of mutual intelligibility, Townend also argues that the English
and Norse were of roughly equal prestige (the second of Hock and Joseph’s
criteria for koinezation):

My suspicion is that Norse and English were roughly adstratal in Viking Age
England – that is, they enjoyed more less equal prestige. For if Old English were of
 The Middle English creolization hypothesis 125

much greater prestige one would expect the rapid death of Old Norse and few Norse
loans in Old English; and if Old Norse were of much greater prestige, one would
expect many Norse loans in Old English of a non-need nature, and certainly not the
death of Old Norse. (1996: 204)

The historical contact setting is therefore much better suited to koinezation than
creolization, though of course the two processes share similarities. Siegel (2001)
makes a convincing argument that creoles and koinés, while differing
significantly from each other in their genesis, structure, and social context, fall at
opposite extremes of a shared continuum of language contact varieties.
McWhorter also sees features of koinezation in Anglo-Norse contact. As a
comparison, he notes that in the koiné Hindustani of Fiji, grammatical features
are often picked “cafeteria style” from the dialects in contact, and that in some
cases the simplification of a feature occurs even where all dialects share the same
form (2005: 306). This situation is plausible in Viking Age England, he asserts,
because complexity in a newly learned language is likely to be lost when there is
interference with a related first language:

A non-native speaker of English, confronted with three forms of the article

corresponding to gender assignments that often conflicted between Old English and
Old Norse, would plausibly have made do with a single gender-neutral marker
rather than applying their native genders to their version of English. Koine data
worldwide… indicate that when structures are cognate but distinct in closely related
varieties, the speaker of one of them is as likely to eliminate their reflex of that
feature as to preserve it even when doing the latter would not appreciably impede
communication (2005: 308).

The accumulated weight of such evidence is enough to make a “koinezation

hypothesis” worthy of further consideration to explain the transition from Old to
Middle English.2
In the above discussion, there is tacit agreement. Advocates and critics of the
creolization hypothesis discuss Middle English as descending from Old English
(the substrate in the creolization hypothesis), with Old Norse and Norman French
having a greater or lesser role in influencing the development of the West
Germanic language. However, Emonds and Faarlund (2014) shift the ancestry of
English to the North Germanic family, arguing that Middle English descended
from the Anglicized Old Norse spoken in the Danelaw, with Old English taking
the role of foreign influencer. This claim falls somewhat outside a discussion of
the creolization hypothesis, but may be seen as a competing contrarian take on

2 Some of the changes that occurred in the development of Middle English, such as the
paradigmatic leveling described here by McWhorter (2005), are parallel to changes in the
transition from Classical to Koiné Greek, the original source of the linguistic term.
126 D. O’Neil

the available evidence. Actually, when the authors do mention the hypothesis
(mainly in footnotes), it is with something like contempt. Clearly, they see the
creolization debate as resulting from scholars being unwilling to consider the
possibility of Middle English as a North Germanic language:

The oddity of Middle English “borrowing” from Scandinavian has spawned a mini-
industry promoting the idea that Middle English is a creole. The maxim for this
school of thought seems to be “we must be cautious, as historically, anything can
happen. That is, there are no predictive principles of language change, and Middle
English proves it.” Such reasoning is obviously circular. (2014: 27)

Emonds and Faarlund base their conclusions on an examination of a broad range

of syntactic evidence (more than 20 constructions), making the reasonable point
that lexical evidence is of less value in determining linguistic inheritance (after
all, Modern English would be a Romance language if familial membership were
determined by counting lexemes). While the argumentation of these authors finds
a degree of approval (see, for example, the review by Lightfoot 2016), their
syntactic evidence is answered with devastating force by Bech and Walkden
(2016), who also point out Emonds and Faarlund’s neglect of phonological
evidence, and by Stenbrenden (2016), who shows their argumentation to be
anachronistic – that is, it uses grammatical features as evidence without attention
to when those features are attested. Furthermore, it should be added that some of
the evidence against creolization (summarized above), such as late West Saxon
case syncretism and the increasing frequency of functors, also serves as evidence
against Emonds and Faarlund. These markers are early signs of English’s
syntactic transition to analytic syntax, strengthening the standard claim that Old
English is the ancestor of Middle English. Emonds and Faarlund argue that “the
morphological simplification was mediated by language contact, and the
simplified inflection cannot in itself be used as an argument for either position”
(2014: 150). However, this would only be the case if the simplification were not
evident in West Saxon during the late Old English period.
Again, Emonds and Faarlund are not themselves proponents of a creolization
hypothesis for Middle English, but their claim of North Germanic ancestry for
the language, with its headline grabbing power (at least in Scandinavia), hints at
another aspect of what makes the creolization hypothesis so enduring. Issues of
language and identity are always salient, but English presents a special case. The
language dominates multiple spheres of international communication, and there
are more combined L1 and L2 speakers of English than there have been for any
language in the history of the world (Crystal 1997). People have strong feelings,
both positive and negative, about the language and its empire. But what if the
stories that English speakers have told themselves about the English language are
not true? What if Vikings (or some other group or groups) spoke the aboriginal
 The Middle English creolization hypothesis 127

language rather than the revered Anglo-Saxons associated with Beowulf, Bede,
and King Alfred the Great? Such theories challenge not only linguistic
dendrograms, but mythologies of linguistic identity that have been passed down
for centuries and become deeply embedded in educational curricula and national
self-esteem.

5. Creolization and Universal Grammar

Even without this high-stakes context, however, it should be clear that identifying
an historical language as a creole is a difficult task. This is especially true when
we insist on the strict definition that a creole is a nativized pidgin. One must trace
a language’s history back to a pidgin stage in order to confirm a diachronic
relationship that began during a time of language contact. For the case of Middle
English, Görlach (1986) argues not only that we have no evidence of a pidgin
stage, but that Old English and Old Norse were too similar for pidginization even
to have occurred. But how could such a claim be tested empirically? Are there
synchronic criteria for identifying which languages are creoles even without
observing the creolization process in the historical record?
Markey (1982) identifies a set of such criteria for classifying Afrikaans, but
McWhorter notes that there has generally been a reluctance among linguists to
think in terms of special creole-specific traits because of concerns that such a list
might be ill-construed, as though these traits mark creoles as inferior languages.
After all, he points out, linguists often feel compelled to demonstrate that all
languages are equal, and a set of distinguishing characteristics that mark creoles
as exceptional might make the language appear somehow unequal – an “other”
among the world’s languages. Milroy (2001b) makes the same point, noting that
in many cases our scholarly declarations on these matters may be directed by
ideological considerations. In fact, he identifies the exact same a priori belief as
McWhorter – namely, that all languages are fundamentally equal – as a certainty
which is ardently held despite the fact that it cannot be demonstrated empirically
(2001b: 624). This is a conflict that I believe has underlain much of the debate
over the creolization hypothesis and the interest it has generated. If English is
understood as a creole, then the “othering” of creoles by their linguistic
characteristics is perhaps not so problematic. Moreover, even if English is not
understood as a creole, the debate sheds light on issues that otherwise might be
ignored. The question persists: Can we classify languages into different groups
on the basis of criteria such as grammatical complexity without attaching value
judgments of superiority and inferiority?
McWhorter himself recalls the cognitive dissonance he experienced when he
was asked by a student about whether creoles could be identified outside their
historical context. Like most linguists, he assumed as an “article of faith” that
128 D. O’Neil

variable complexity between two languages in particular aspects of their structure

should “balance out” once every aspect had been examined. In other words, if a
language were found to have little complexity in one area of its grammar (say,
nominal inflections), then there should be a higher level of complexity elsewhere,
leading to equivalence between all languages in terms of their complete
description. This non-empirical assumption seems supported by Chomsky’s
theory of Universal Grammar (UG), which relies on the claim that all forms of
languages are reflections of an underlying linguistic apparatus based in the human
genetic code.

The concept of UG initiated by Chomsky can be defined as the scientific theory of

the genetic component of the language faculty… It is the theory of that feature of
the genetically given human cognitive capacity which makes language possible,
and at the same time defines a possible human language. UG can be thought of as
providing an intensional definition of a possible human language, or more precisely
a possible human grammar…. (Roberts 2016: 2)

UG has come under increasing criticism over the last decade (e.g., Christiansen
& Chater 2009, Evans & Levinson 2009, Tomasello 2009, Dąbrowska 2015,
among others), but Chomsky’s theory has arguably been “the predominant
approach in linguistics for almost 50 years” (Dąbrowska 2015: 1). One
implication of the theory is that all languages should be governed, at least at some
level, by the same grammatical structure. Although this grammar would not
necessarily need to be expressed in the same degree of complexity in every
language, many linguists have at least had the feeling that it should be so, as
exemplified in McWhorter’s anecdote (and, I might add, my own experience).
Yet McWhorter was troubled by his response to the student, which he later
decided had been doctrinaire and unreflective. He returned to the matter and
concluded that creoles do share certain structural characteristics with each other
that distinguish them from other languages.
The formalization of this discovery is called the Creole Prototype hypothesis.
McWhorter was not the first to consider formal criteria for identifying creoles
outside of a clear historical context (again, see Markey 1982 for criteria used to
classify Afrikaans), but McWhorter’s hypothesis is notable for its efficiency. He
finds three features that all creoles lack: inflectional morphology, contrastive tone
(such as in Chinese, where pitch alone is enough to distinguish lexemes), and
noncompositional derivational morphology (as in the English understand). These
three features are examples of what a person would often struggle with when
learning a language without formal schooling. More generally, McWhorter
argues that creoles should exhibit less “overspecification” (unnecessary
complexity such as gender and various verbal markers), because this sort of
complexity, which is unnecessary for communication, only accrues over
 The Middle English creolization hypothesis 129

generations. The point is not to rank languages by complexity, but to identify

languages that lack the kind of unnecessary complexity that could not have
emerged in the first generations following creolization. Arguably, English
exhibits less complexity and overspecification than any other Germanic language,
even Afrikaans, a “semi-creole” descendant of Dutch, which is a cousin of
English in the West Germanic language family. Nevertheless, by the rather strict
definition of the Creole Prototype hypothesis, Middle English would still not be
considered a creole, because it exhibits only two of the three required features.
This conclusion rather strikingly affirms Domingue’s rather tentative take on the
creolization hypothesis, which concluded with the claim that Middle English is
perhaps best described as “very much alike a creole” (1977: 89).
Despite his own initial state of uneasiness, McWhorter’s hypothesis has little
in common with creole theories that arose in the nineteenth-century milieu of
racism and linguistic ethnocentrism. Differences in complexity notwithstanding,
a creole is obviously just as useful for human communication as any other
language, and questions of linguistic inferiority are really beside the point. This
position has been stated repeatedly by modern linguists – for example, by Poussa,
who asserts creoles have full expressive power even if they may be less fully
developed in some respects than other languages (1982: 70) – yet some linguists
feel that theories such as the Creole Prototype hypothesis represent a throwback
to reactionary views of language. DeGraff, for example, is unflinching in his
condemnation of any theory that might somehow segregate creoles from other
languages, asserting that “the Creole prototype has actually little to contribute to
our insights into the definitely-not-so-simple mechanics of universal grammar
and no relevance to current debates about the correct analyses of an array of
complex Creole structures” (2003: 400). But DeGraff also takes the rather
extreme position that pidgins and creoles do not actually exist (which recalls
Bailey and Maroldt in their original framing of the creolization hypothesis), and
he argues that it is erroneous and prejudicial to view the linguistic processes that
form them as exceptional.

6. Conclusion

To conclude, it should be noted that creolization, understood as the creative act

of a generation of speakers growing up with a pidgin, entails no pejorative
judgment. In fact, it would only be possible for language learners to convert a
rudimentary communication system such as a pidgin into a full-fledged language
in a single generation if their cognitive apparatus was inherently suited to the
complex phonological, syntactic, and semantic rules that the creolization process
requires. In this view, the fact of creolization is a testament to the astonishing
linguistic creativity of creole speakers, and modern creolization theories should
130 D. O’Neil

not be mistaken as having anything in common with the colonialist absurdities of

nineteenth-century linguistic theory. Yet it is unsurprising that theories of
creolization become particularly controversial when applied to a case like Middle
English. More is at stake in such a discussion than might first meet the eye – in
our understanding of creole theory, in our political views on English and other
standard world languages, and in even in theories of Universal Grammar and the
standard inheritance model of historical linguistics. The scholarly question of
whether English is a creole is thus of inherent rhetorical interest and worthy of
continued discussion, even if the evidence (as reviewed above) is mostly one-
sided. Sometimes the merit of asking a question is not just in its answer, but in
the issues that question encourages us to explore along the way.

REFERENCES

Allen, Cynthia. 1997. Middle English case loss and the ‘creolization’ hypothesis. English Language
and Linguistics 1.1. 63–89. DOI: 10.1017/S1360674300000368
Bailey, Charles James & Karl Maroldt. 1977. The French lineage of English. In Jürgen Meisel (ed.),
Langues en contact: Pidgins, creoles, Tübingen: TBL. 21–53.
Bech, Kristin & George Walkden. 2016. English is (still) a West Germanic language. Nordic
Journal of Linguistics 39.1. 65–100. DOI: 10.1017/S0332586515000219
Christiansen, Morten H. & Nick Chater. 2009. The myth of language universals and the myth of
universal grammar. Behavioral and Brain Sciences 32.5. 452–453. DOI:
10.1017/S0140525X09990641
Crystal, David. 1997. English as a global language. Cambridge: Cambridge University Press. DOI:
10.1017/CBO9780511486999
Dalton-Puffer, Christiane. 1995. Middle English is a creole and its opposite: On the value of
plausible speculation. In Jacek Fisiak (ed.), Linguistic change under contact conditions,
Berlin: Mouton de Gruyter. 35–50. DOI: 10.1515/9783110885170.35
Danchev, Andrei. 1997. The Middle English creolization hypothesis revisited. In Jacek Fisiak,
(ed.), Studies in Middle English linguistics, Berlin: Mouton de Gruyter. 79–108. DOI:
10.1515/9783110814194.79
DeGraff, Michel. 2003. Against creole exceptionalism. Language 79.2. 391–410.
Domingue, N. Z. 1977. Middle English: Another creole? Journal of Creole Studies 1. 89–100.
Dąbrowska, Ewa. 2015. What exactly is Universal Grammar, and has anyone seen it? Frontiers in
Psychology 6:852. 1–17. DOI: 10.3389/fpsyg.2015.00852
Emonds, Joseph E. & Jan Terje Faarlund. 2014. English: The language of the Vikings. Olomouc:
Palacký University Press.
Evans, Nicholas & Stephen C. Levinson. 2009. The myth of language universals: Language
diversity and its importance for cognitive science. Behavioral and Brain Sciences 32.5.
429–448. DOI: 10.1017/S0140525X0999094X
Görlach, Manfred. 1986. Middle English – a creole? In Dieter Kastovsky & Aleksander Szwedek
(eds.), Linguistics across historical and geographical boundaries, Berlin: Mouton de
Gruyter. 329–344.
 The Middle English creolization hypothesis 131

Hartman, Megan E. 2016. New applications for word-foot theory. In M. J. Toswell & Lindy Brady
(eds.), Early English poetic culture and meter: The influence of G. R. Russom,
Kalamazoo: Medieval Institute Publications. 73–92.
Hock, Hans Henrich & Brian D. Joseph. 1996. Language history, language change, and language
relationship: An introduction to historical and comparative linguistics. Berlin: Walter
de Gruyter & Co.
Hogg, Richard M. 1997. Some remarks on case marking in Old English. Transactions of the
Philological Society 95.1. 95–109. DOI: 10.1111/1467-968X.00014
Ingham, Richard. 2014. The maintenance of French in later medieval England. Neuphilologische
Mitteilungen, 115.4. 425–448.
Lightfoot, David W. 2016. Review of Joseph E. Emonds & Jan Terje Faarlund, English: The
language of the Vikings, 2014. In Language 92.2. 474–477.
Markey, Thomas L. 1982. Afrikaans: Creole or non-creole? Zeitschrift für Dialektologie und
Linguistik, 49.2. 169–207.
McWhorter, John H. 2005. Defining creole. Oxford: Oxford University Press.
McWhorter, John H. 2008. Our magnificent bastard tongue: The untold history of English. New
York. NY: Gotham Books.
Milroy, James. 2001a. Language ideologies and the consequences of standardization. Journal of
Sociolinguistics 5.4. 530–555. DOI: 10.1111/1467-9481.00163
Milroy, James. 2001b. Response to Sally Johnson: Misunderstanding language? Journal of
Sociolinguistics 5(4). 620–625. DOI: 10.1111/1467-9481.00170
O’Neil, David. 2017. A syntactic basis for the distribution of metrical types in Beowulf. The
Mediaeval Journal 7.1.29–59. DOI: 10.1484/J.TMJ.5.115346
O’Neil, David. 2018. English as the lingua franca of international publishing. World Englishes 37.2.
146–165. DOI: 10.1111/weng.12293
O’Neil, David. 2020, forthcoming. Syntax, prosody, and the Brut: Metrical continuity in the early
English alliterative tradition. Enarratio 23.
Plag, Ingo. 1998. Review of Christiane Dalton-Puffer, The French influence on Middle English
morphology: A corpus-based study of derivation, 1996. In Language 74.2. 392–395.
Phillipson, Robert. 1992. Linguistic imperialism. Oxford: Oxford University Press.
Poussa, Patricia. 1982. The evolution of early standard English: The creolization hypothesis. Studia
Anglia Posnaniensia 14. 69–85.
Roberts, Ian. 2016. Introduction. In Ian Roberts (ed.), The Oxford handbook of Universal Grammar.
Oxford: Oxford University Press. DOI: 10.1093/oxfordhb/9780199573776.013.1
Russom, Geoffrey R. 2017. The evolution of verse structure in Old and Middle English poetry:
From the earliest alliterative poems to iambic pentameter. Cambridge: Cambridge
University Press. DOI: 10.1017/9781316562925
Siegel, Jeff. 2001. Koine formation and creole genesis. In Norval Smith & Tonjes Veenstra (eds.),
Creolization and contact, Amsterdam: Benjamins.175–197. DOI: 10.1075/cll.23.08sie
Stenbrenden, Gjertrud F. 2016. Why English is not dead: A rejoinder to Emonds and Faarlund.
Folia Linguistica Historica 37.1. 239–279. DOI: 10.1515/flih-2016-0008
Tomasello, Michael. 2009. Universal grammar is dead. Behavioral and Brain Sciences 32.5. 470–
71. DOI: 10.1017/S0140525X09990744
Thomason, Sarah Grey & Terrence Kaufman. 1988. Language contact, creolization, and genetic
linguistics. Berkeley, CA: University of California Press.
Townend, Matthew. 2002. Language and history in Viking Age England: Linguistic relations
between speakers of Old Norse and Old English. Turnhout: Brepols.
 132 D. O’Neil

Trotter, David M. 2017. Middle English creolization. In Laurel J. Brinton & Alexander Bergs (eds.),
Middle English, Berlin: Mouton de Gruyter. 224–238. DOI: 10.1515/9783110525328-
012
Watts, Richard J. 2011. Language myths and the history of English. Oxford: Oxford University
Press. DOI: 10.1093/acprof:oso/9780195327601.001.0001

View publication stats