BORROWING
BORROWING
BORROWING
Jeanine Treffers-Daller
Introduction
Constraints
Quantitative approaches
Psycholinguistic approaches
(please note the date above is the from the previous version of this article see also
correspondence with Jef Verschueren)
Introduction
Borrowing - the incorporation of features of one language into another - has been
studied by researchers working in a wide range of areas, from a diachronic as well
as a synchronic perspective. In a diachronic research tradition that focuses on the
historical development and the genetic classification of languages, it is clearly of
central importance to be able to distinguish borrowed features from non-borrowed
or native features of a language. The importance of the role of lexical and
structural borrowing for language change has however only fairly recently been
recognised, as until the publication of Thomason and Kaufmans (1988) book on
contact-induced change, historical linguists used to emphasise internal causes of
language change. Interest in the synchronic analysis of borrowing emerged
towards the turn of the century. The main focus of the synchronic analyses has
been to identify the grammatical constraints on borrowing, and to describe the
phonological, syntactic and morphological integration of borrowed words. In
addition, researchers have tried to delimit borrowing from other language contact
phenomena, such as code-switching and transfer, and they have developed
different classifications of borrowing. The social correlates of borrowing have
received attention in more quantitatively oriented studies. Put differently,
researchers have mainly focused on what Weinreich, Herzog and Labov (1968)
have called the embedding problem and the constraints problem. The embedding
problem, when applied to the study of borrowing, concerns on the one hand the
embedding or integration of source language features in the borrowing language.
On the other hand, it deals with the embedding of these features in the social
structure: to what extent do social factors influence the quantity and the quality
of the borrowing process. The constraints problem deals with the question of
determining the set of possible borrowings and with the discovery of the
structural constraints on the borrowing process. Other questions have received
less systematic attention. The actuation problem and the transition problem (how
and when do borrowed features enter the borrowing language and how do they
spread through the system and among different groups of borrowing language
speakers) have only recently been studied. The evaluation problem (the
subjective evaluation of borrowing by different speaker groups) has not been
investigated in much detail, even though many researchers report that borrowing
is evaluated negatively. Apart from the issues raised above, in more recent
studies, pragmatic and psycholinguistic aspects of borrowing have been studied in
some detail.
Milroy (1997: 311) points out that as for language contact, it is not actually
languages that are in contact, but the speakers of the languages. I agree with
Milroy that there is a danger in seeing languages as discrete entities independent
of speakers, because under this view the role of speakers in actuating and
diffusing language change (or borrowing) is being neglected. Most researchers
have however continued to use the terms borrowing, albeit in slightly differing
definitions.
Researchers have different views on the necessity to distinguish borrowing
from other language contact phenomena, such as transfer, convergence and
code-switching. Poplack and associates (Poplack 1980; Poplack and Meechan
1995) maintain that borrowing and code-switching are different phenomena.
Support for this position comes from Grosjean and associates ( Grosjean 1988,
1995, 1997), who have extensively studied psycholinguistic aspects of codeswitching and borrowing (see under psycholinguistic approaches for more details).
Other researchers assume that there is a common set of formal principles to
morphological and syntactic structure and that as a result, there may well be
parallel constraints on borrowing and code-switching (Appel and Muysken 1987;
Muysken 1990). The issue cannot be discussed here in more detail (cf. TreffersDaller 2005; 2009).
lexical
interference),
but
also
syntactic
and
phonological
to
borrowing
developed
before.
The
sociolinguistic
and
psycholinguistic aspects of his work are now out of date, but the classifications of
different types of borrowing Weinreich and Haugen developed are still widely
used. As they are very similar, the following discussion is limited to Haugen's
typology.
Haugen (1950) distinguishes different types of borrowing, based on the
question of whether or not source language morphemes are imported into the
borrowing and whether or not substitution of source language morphemes or
phonemes by borrowing language elements occurs (see alsoBackus and Dorleijn
2009 for a further refinement of this typology). We will illustrate the concepts
with examples from French-Dutch language contact data, as described in TreffersDaller (1994).
When French-Dutch bilinguals import the French discourse marker donc (so) into
Brussels Dutch, they keep the French nasal vowel [] and do not replace it with
one or more Dutch phonemes.
In the case of loan shifts only a meaning, simple or composite, is imported, but
the forms representing this meaning are native (Appel and Muysken 1987: 165).
Famous examples are German Wolkenkratzer, French gratte-ciel and Spanish
rascacielos, all of which are modelled on English skyscraper (Haugen 1950: 214).
In these cases both halves of the compound have been translated into the
borrowing language, but no source language morphemes have been imported into
the borrowing language. Similarly, in some cases, the meaning of a word can be
extended or changed without any importation of lexical material. An example of
this phenomenon, called semantic loan, is found in the extension of the meaning
of Dutch tellen (to count). In Belgian Dutch, tellen has acquired the meaning to
count on from French compter sur, as in the expression: iemand waar ge op kunt
tellen (someone you can count on).
Poplack and associates introduced a different typology of borrowing, based
on the diffusion of these elements throughout a speech community:
Established loanwords (which typically show full linguistic integration, nativelanguage synonym displacement and widespread diffusion, even among recipientlanguage monolinguals) differ from nonce borrowings only insofar as the latter
need not satisfy the diffusion requirement ( Poplack and Meechan 1995: 200).
phonemes. Many speakers will substitute [] with native [u], which is an example
of adaptation in Van Coetsem's terminology.
Haugen (1950: 222); Poplack, Sankoff and Miller (1988) and Thomason and
Kaufman (1988) show that there is a lot of variability in the pronunciation of
loanwords and that this variation correlates with speaker characteristics such as
age and bilingual ability. Older speakers who have a less elaborate command of
the source language phonology integrate the loanwords to a larger extent into the
phonological patterns of the borrowing language than younger speakers do.
Poplack and Sankoff (1984) and Poplack, Miller and Sankoff (1988) provide
evidence for the fact that phonological integration proceeds as a function of the
social integration of the loanword. Widespread loans which have entered the
borrowing language at an early date are often realised with borrowing language
pronunciation, whereas the pronunciation of more recent and less widespread
ones is often more similar to the pronunciation in the source language.
Thomason and Kaufman (1988: 124) demonstrate that the importation of
large numbers of loanwords into a language does not necessarily have important
consequences for the phonological system of the borrowing language. Although
lexical influence of French on English was very heavy, there is very little structural
interference from French. French loanwords did not introduce any new phones at
all into English, according to Thomason and Kaufman, even though formerly
allophonic distinctions, such as the distinction between [f] and [v], were
phonemicized in Middle English under the influence of French. In other language
contact situations the phonological system of the borrowing language can be
changed dramatically, as the case of Asia Minor Greek (Dawkins 1916, in
Thomason and Kaufman 1988) illustrates.
As far as the syntactic integration of loanwords is concerned, gender
allocation to borrowed nouns is a well explored area (Baetens Beardsmore 1971;
Chirsheva 2009; Poplack et al 1982; Wawrzyniak 1985; Poplack, Sankoff and
Miller 1988). Morphological integration has been studied by Miller (1997), who
focuses on the combination of French derivational suffixes and English roots and
vice versa. Others have investigated the addition of inflectional morphology, such
as the formation of plurals of borrowed words (Poplack et al 1988; Treffers-Daller
1999).
Constraints
The main thrust of research at the end of the seventies and the eighties was to
discover the constraints on the borrowing process. It was clear that most bilingual
data contained loanwords of different categories and that some categories were
more likely to be borrowed than others. As mentioned above, all researchers
recognise the existence of lexical borrowing and this is certainly the wide-spread
form of borrowing in the languages of the world. As Weinreich (1953: 56) puts it,
the vocabulary of a language, considerably more loosely structured than its
phonemics or its grammar, is beyond question the domain of borrowing par
excellence. In addition, Weinreich points to the socio-cultural reasons behind the
fact that the lexicon is so receptive to borrowing: speakers often want to
introduce new concepts or try to avoid homonyms and frequently replace outworn
expressions with new ones.
The aim of the studies into constraints on borrowing was to explain why
certain categories were more likely to be borrowed than others. Appel and
Muysken (1987: 170171) give an overview of different approaches to this
problem,
and
discuss
the
principles
behind
the
so-called
hierarchies
of
Muysken (1981a) formulates the following hierarchy on the basis of his analysis of
Spanish borrowings in Quechua:
(2
Appel and Muysken (1987: 172) explain that paradigmatic and syntagmatic
coherence relations in language are at the basis of these hierarchies. As the
pronoun system of a language is more tightly organised than the adjectives, for
example, pronouns are less likely to be borrowed. Categories that are firmly
embedded in the syntagmatic relations in a sentence, such as verbs, are less
likely to be borrowed than elements such as nouns, which are less crucial to the
organisation of the sentence.
Moravcsik (1978) formulates the constraints on borrowing in the form of
implicational universals of borrowing. Matras (1998: 283) summarises and
situation can be found in the adoption of French phonemes, but their occurrence
is restricted to loanwords. Thus, the French nasal vowels are imported in some
loan words, such as donc and de temps en temps. Syntactic borrowing is very
limited in Brussels Dutch. Borrowed adverbial particles appear in a pre-clausal
position, rather than in the first position in the main clause and this phenomenon
can perhaps be attributed to influence from French (Treffers-Daller 1994, 1999).
At level three, one finds borrowing of prepositions and slight structural
borrowing. Derivational affixes may be abstracted from borrowed words and
added to native vocabulary. Miller (1997) in his study of borrowed suffixes on
native bases in Middle English, comes to the conclusion that French suffixes were
productive in Middle English, as there were at least 100 hybrid forms consisting of
an English base and a French suffix in English prior to 1450. The most productive
of these were: -able as in understandable, -ess as in murderess, and -ery as in
husbandry.
At levels four and five, major structural features are borrowed. According to
Thomason and Kaufman, Weinreich's (1953) examples of German borrowing in
Romansh, which has lost gender in predicative adjectives and in which the nounadjective word order is partially replaced by adjective-noun order, is probably an
example of borrowing at level four. Thomason and Kaufman extensively discuss
Dawkins' (1916) analysis of Turkish influence on Asia Minor Greek, which they see
as an example of structural borrowing at level five. In these varieties of Greek,
various word order features are borrowed, as well as vowel harmony. In addition,
several grammatical categories that Turkish lacks, such as gender and adjectivenoun agreement are lost.
Bakker and Mous (1994) and Thomason (1996) convincingly show that
there is a difference between languages which are characterised by heavy
borrowing and genuinely mixed languages. Bakker and Mous define mixed
languages as a combination of the grammatical system of one language with the
lexicon of another. An example is Media Lengua, which is a combination of
Quechua grammar with a Spanish lexicon ( Muysken 1981b, 1994, 1996). Michif
is a different case, because it has the nominal grammar from French and the
verbal grammar from Cree (Bakker 1996). English does not classify as a mixed
language, despite the fact that 75% of its vocabulary is from French, because the
basic vocabulary of English remains almost completely English. According to
Thomason and Kaufman (1988) only 7% of the basic vocabulary of English is
borrowed. In mixed languages such as Media Lengua, almost all words have been
replaced, including the basic vocabulary.
As borrowing and interference through shift can occur in the same contact
situation, it is sometimes difficult to establish with certainty whether the
phenomena under consideration are due to one or the other process. Especially in
those
situations
characterised
by
heavy
borrowing
or
heavy
substratum
Romani, Matras shows that a number of pragmatic principles explain the different
degrees of borrowability of utterance modifiers which have the same structuralsyntactic status. Instead of aiming for a structural syntactic approach to a
hierarchy of borrowability, Matras develops a function-related hierarchy in which,
for example, contrastive conjunctions are very frequently borrowed, but temporal
conjunctions much less frequently. Matras argues that the process responsible for
the replacement of the utterance modifiers is fusion, i.e., the nonsepararation of
the two systems, and he points to psycholinguistic factors, such as the cognitive
pressure exerted on bilinguals to draw on the resources of the pragmatically
dominant language for situative, gesturelike discourse-regulating purposes
(Matras 1998: 321) to explain the hierarchy. In a follow-up study Matras (2000)
develops the cognitive model for bilingual discourse markers further and shows
how bilingual speakers accidentally produce discourse markers from language B
while aiming to speak language A, which is interpreted as evidence for the
existence of fusion.
Although pragmatic and discourse-functional perspectives on borrowing
remain rare, Prince (1988) is one of the very few studies which address the issue
of pragmatic borrowing, i.e., the borrowing of a discourse function of a particular
syntactic form from another language. Prince shows that the discourse function of
Yiddish DOS-sentences, exemplified in (3), was borrowed from Russian, on the
analogy of the Slavic expletive ETO this:
(3)
Dos
shlogst
du
Di
puter?
This
beats
you
The
butter
are
functional
equivalents
of
English
cleft-sentences,
but
Es
dremlen
feygl
oyf
Di
tsvaygn
it
doze
birds
on
The
branches
Quantitative approaches
by
verbs,
adjectives
and
conjunctions.
Interjections
and
frozen
less than 1% of the total words in the corpus, showing that borrowing is a
relatively rare phenomenon in this data base. Treffers-Daller (1994) found similar
low rates of borrowing in a French-Dutch corpus of 190.000 words and showed
that there was a clear asymmetry in the directionality of borrowing: French words
were far more likely to be borrowed into Dutch than Dutch words into French.
Using advanced statistical techniques, van Hout and Muysken (1994) study
the influence of four sets of factors on borrowability in a Spanish-Quechua
bilingual corpus: lexical content, frequency, structural coherence factors and
equivalence.
They
show
that
structural
coherence
factors,
in
particular
Psycholinguistic approaches
clear that important new insights into the characteristics of borrowing and the
differences between borrowing and other language contact phenomena can be
gained from psycholinguistic approaches. Grosjean (1985 et seq) introduced the
concept of language modes, which has subsequently been shown to be a very
powerful explanatory concept:
Bilinguals find themselves in their everyday lives at various points along a
situational continuum that induce different language modes. At one end of the
continuum, bilinguals are in a totally monolingual language mode, in that they are
interacting with monolinguals of one or the other of the languages they know.
At the other end of the continuum, bilinguals find themselves in a bilingual mode,
in that they are communicating with bilinguals who share their two (or more)
languages and with whom they normally mix languages (i.e., code-switch and
borrow) (Grosjean 1997: 228).
When researchers control for the language mode their subjects are in, they have
a better chance of disentangling the different language contact phenomena found
in the speech of bilinguals, and this may in turn help to identify the (differences
and similarities between) the constraints on the various phenomena.
Whereas in a bilingual language mode, all bilingual language phenomena can
occur (interferences, code-switches, borrowings of various types, etc.) this is not
the case in a monolingual language mode. Here code-switches and borrowings
are either inexistent or are usually kept to a strict minimum so as to ensure
adequate communication (Grosjean 1998: 228).
Grosjean (1997) also discusses evidence for the fact that borrowing and codeswitching are processed differently.
The concept of language modes is one of the basic characteristics of
Grosjean's interactive activation model of word recognition in bilinguals, named
BIMOLA (Bilingual Mode of Lexical Access), developed over many years of
experimental research (see Grosjean 1997 for a detailed presentation).
Green's (1998) Inhibitory Control Model is a model of bilingual speech
processing which aims at explaining how bilinguals control their two languages,
for example when translating from L 1 to L2 without actually using L1 words. This is
done by assuming that lemmas are specified in terms of a language tag. Thus,
each lemma has an associated language tag and this tag is one of the factors
which affects the activation of the lemma. After lemmas have been linked to
lexical concepts, the model allows for lemmas with the wrong tag to be
inhibited, so that they cannot catch speech production during a translation task.
Though the model was not developed to account for borrowing and code-
switching, it may well have interesting implications for the analysis of language
contact phenomena. One of the issues to be investigated further is whether the
concept of language tag(s) attached to the lemma's of words can be helpful in
distinguishing established borrowings, nonce borrowings and code-switches. It is
well known that many established borrowings can no longer be recognized as
borrowings. Thus, native speakers of English do not recognize people as a
borrowing from French, even if they are fluent speakers of French. Guest words
that contain strong language phonetic or phonotactic cues, such as snob in
French, are probably still recognizable. One may wonder whether in the process
of borrowing language tags are lost or replaced and whether language tags play a
role below or beyond the word level (Treffers-Daller 1998). As detailed phonetic
analyses of individual segments and suprasegmental features are now possible
with the help of software such as PRAAT (http://www.fon.hum.uva.nl/praat/),
new insights into the phonetic characteristics of borrowings and other language
contact phenomena may well become available in the near future, and these may
help to establish whether it is necessary and feasible to distinguish different
contact phenomena from each other either receptively or productively.
BLOM, J.-P. & J.J. GUMPERZ (1972) Social meaning in linguistic structure: codeswitching in Norway. In J.J. Gumperz & D. Hymes (eds.) Directions in
sociolinguistics: 407434. Holt, Reinhart & Winston.
BLOOMFIELD, L. (1933) Language. Holt, Reinhart & Winston.
CHIRSHEVA, G. (2009). Gender in Russian-English code-switching. International
Journal of Bilingualism 13 (1), 20-93.
CLYNE, M. (1967) Transference and triggering. The Hague.
CLYNE, M. (2003) Dynamics of language contact. Cambridge: CUP.
COLANTONI, L. & J. GURLEKIAN (2004) Convergence and intonation: Historical
evidence from Buenos Aires Spanish. Bilingualism: Language and Cognition 7(2):
107119.
DAWKINS, R.M. (1916) Modern Greek in Asia Minor: a study of the dialects of Slli,
Cappadocia and Phrasa with grammars, texts, translations and glossary.
Cambridge University Press.
DE ROOIJ, V.A. (1996) Cohesion through contrast: Discourse structure in Shaba
Swahili/French conversations. IFOTT.
EXTRA, G. & L. VERHOEVEN (eds.) (1993) Immigrant languages in Europe. Multilingual
Matters.
FULLER, J.M. (2001) The principle of pragmatic detachability in borrowing: Englishorigin discourse markers in Pennsylvania German. Linguistics 39(2): 351369.
GREEN, D. (1998) Mental control of the bilingual lexico-semantic system. Bilingualism:
Language and Cognition 1(2): 6781.
GROOT, A. de & J.F. KROLL (eds.) (1997) Tutorials in Bilingualism. Psycholinguistic
perspectives. Lawrence Erlbaum.
GROSJEAN, F. (1985) The bilingual as a competent but specific speakerhearer. Journal
of Multicultural and Multilingual Development 6: 467477.
(1988) Exploring the recognition of guest words in bilingual speech. Language and
Cognitive Processes 3: 233274.
(1995) A psycholinguistic approach to code-switching: the recognition of guest
words by bilinguals. In L. Milroy & P. Muysken (eds.): 259275.
(1997) Processing mixed language: issues, findings, and models. In A. de Groot &
J.F. Kroll (eds.) Tutorials in Bilingualism. Psycholinguistic perspectives: 225254.
Lawrence Erlbaum.
GROSJEAN, F. & C. SOARES (1986) Processing mixed language: some preliminary
findings. In J. Vaid (ed.) Language processing in bilinguals: psycholinguistic and
neuro-psychological perspectives: 145179. Lawrence Erlbaum.
Haspelmath, M. and U. Tadmor (2009) Loanwords in the worlds languages: a
comparative handbook. Berlin: Mouton de Gruyter.