This document discusses etymology and pidgin and creole languages. It notes that pidgin and creole languages provide an opportunity to develop more sophisticated models of etymological research due to the complex mixing at the lexical level in these languages. The document then examines the multiple origins of lexical material in pidgin languages, using Tok Pisin as a case study. It provides examples of Tok Pisin words that have origins in both English and local languages, or in English, German and Malay. The document argues that a significant portion, perhaps 25%, of early Tok Pisin words have multiple etymologies due to the contact between multiple languages in the formation of pidgins.

1. ETYMOLOGY AND PIDGIN AND
CREOLE LANGUAGES
By PETER MUHLHAUSLER
1. INTRODUCTION
In a linguistic framework of description where synchronic
investi-
gation is regarded as methodologically prior to diachronic
investi-
gation (i.e. the prevailing paradigm derived from Saussure and
Chomsky), there is little room for etymological studies. The
decline
of such studies has been recently discussed by Malkiel (1975:
101-
120) and proposals were put forward by him to restore
etymologi-
cal research to a position nearer to the core of linguistics. With
the
renewed emphasis, in the most recent past, on developmental
and
historical aspects of language (e.g. Bailey, 1980) there is hope
that a
reassessment of this neglected sub-field of linguistics is
imminent. It
is likely that pidgin and creole languages, where mixing at the
lexical level is particularly intricate, will provide the point of
depar-
ture for more sophisticated models of etymological research.
The main arguments against an etymological approach to

2. language include :
(i) that most researchers are misled by the ‘etymological
fallacy’,
i.e. the belief that the meaning of words can be determined by
investigating their origin;
(ii) that it hinges on chance discoveries, flashes of imagination
and
accident;
(iii) that it is a time-consuming process yielding few insights
rele-
vant to other areas of linguistics.
I feel that the only criticism that stands up to closer inspection
is
that etymologizing remains a very time-consuming business. All
other weaknesses can be mended and I d o not see why a well
developed theory of etymology could not provide vital
information
to researchers in many areas of language change. There will no
100 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
doubt need to be some rethinking, in particular in the
development
of methods which are sensitive to the fact that many languages
undergo processes of linguistic change other than those which
can
be accommodated within a family-tree model of linguistic
relation-
ships. One group of such languages are pidgins and Creoles.

3. Wood
claims for them that
The methods of classical etymology . . . are not directly
applicable to
non-conventional languages such as Creoles . . .
(R. Wood, 1972 quoted from Edwards, 1974 : 5 )
Other languages which may provide similar challenges to
tradition-
al methods of etymologizing are discussed by Hockett (1950).
It remains to be seen whether pidgins, Creoles and other non-
traditional languages require qualitatively different new
approaches
to etymologizing or whether they are merely characterized by a
greater concentration of problems well known to those who
have
worked on more traditional languages.
As the terms pidgin and creole are used in many different senses
it would seem advisable briefly to define them for this paper:
By a p i d g i n one understands functionally and structurally
re-
duced languages, with no native speakers, used to communicate
across linguistic boundaries. Pidgins typically begin as drasti-
cally impoverished languages (called jargons) but can become
highly elaborate over a long period of time. For a pidgin to
stabilize and develop at least three languages must be involved.
Creoles are pidgins which have acquired a community of native
speakers. They typically develop among children whose parents
do not share the same vernacular and therefore have to com-
municate in a pidgin. In the course of creolization pidgins are
made functionally and structurally comparable to other full
languages.

4. Although most of my examples have been taken from Tok Pisin
(Pidgin English of Papua New Guinea), the problems I shall
address myself to are not language-specific. I have chosen to
limit
my discussion to the following points:
(i) multiple origin of lexical material
(ii) discontinuous developments
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 101
(iii) the ‘etymological fallacy’ fallacy
(iv) lexical creativity
(v) practical etymologizing for Tok Pisin.
2. MULTIPLE ORIGIN OF LEXICAL MATERIAL
As pointed out by a number of writers (e.g. Whinnom, 1971),
pidgin
languages result from contact between three or more languages.
The linguistic outcome, it is widely believed, is a pidgin which
has
the lexicon of one language (the ‘lexifier’ language spoken by
the
socially dominant group) and the grammatical structure of one
or
more other languages. (These are assumed to be substratum
languages in earlier, universal grammar in more recent
accounts.)
The alleged lexical purity of various pidgins and Creoles can be
seen from the following figures based on Hall (1966), Bollke
(1980),

5. Laycock (1970) and Mihalic (1971):
Seychellois: French 97%, other 3%
Papiamento: Spanish 80%, other 20%
Tok Pisin: English 79%, Tolai 11 %, other New Guinea
languages 6%, German 3%, Malay 1%
Note that these figures refer to lexical types not tokens.
These figures are rather surprising given that, in the case of Tok
Pisin for instance, speakers of many languages were in contact
in
its formative years and that, in early culture contact, the social
dominance of the English-speaking Europeans was a very shaky
one. Equally surprising is the absence of figures on shared
lexical
items. My investigations into the origin of the Tok Pisin Iexicon
seem to indicate that the extent of shared lexical items is quite
significant, probably as high as 25% in the formative years of
the
language. Let us consider an actual example: two recent diction-
aries, Steinbauer (1969) and Mihalic (1971), give the origin of
bel
‘belly, stomach, seat of emotions’ as (E) = English. A look at
Tok
Pisin’s most important substratum language, Tolai, reveals the
existence of a lexical item bala meaning ‘stomach, seat of
emotions’
in this language. There would seem to be good reasons for as-
suming that be1 and its variant bele were selected by members
of
102 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY

6. 1982
both groups involved in the initial contact situation because it
could be identified across both languages.’ Similar observations
have been made for other languages. Thus, Le Page (1974 : 49)
remarks in connection with the origin of West African Pidgin
Eng-
lish :
Contact situations are bound to involve a good deal of
exploration by
both speaker and hearer, which will inevitably result in some
lucky and
many fruitless sallies. The lucky ones are likely to be
immediately rein-
forced by the participants, each eager to snatch at means of
communi-
cation; the unlucky ones are unlikely to be often repeated.
Coincidence
of form with some similarity of meaning between items from
two codes
will mean that such items will have a high probability of
survival in the
emergent pidgin code. A lexical example would be English dirty
and Twi
doti jointly giving rise to some pidgin fore-runner of Jamaican
creole
dati.
Such cases could be, and in fact have been, shrugged off as
margin-
al phenomena. However, my research into Tok Pisin appears to
point in a different direction, although linguists working on this
language have been slow to admit it. That a number of Tok
Pisin
items can be related to both English and local languages was

7. first
pointed out by Nevermann (1929: 253-4):
Some Pidgin words which at first glance appear to be Engli sh
have,
however, only a chance similarity to it. Thus, the Tolai word
kiap’ ‘chief‘
has nothing to d o with ‘captain’ but is native. Pusi ‘cat’ also
seems not to
be connected with English ‘pussy’ but is probably Samoan. The
word for
’women’, mari or meri, which is usually derived from the name
‘Mary’,
popular among sailors, seems to me to be connected rather with
the
Tolai word mari ‘to love’ or mari ‘pretty, beautiful’, if it is not
to be
derived from married.
(author’s translation)
This quotation clearly illustrates the reluctance of linguists at
the
time to acknowledge the possibility that a lexical item may be
the
result of conflation, and their consequent insistence on single
‘true’
etymologies. The possibility of conflation was acknowledged
later,
however, by Bateson (1944: 138), who argued as follows:
In a few cases, a single word may combine both English and
native
roots. The word liklik, meaning ‘small’, for example, is such a
hybrid
between the English little and ikilik, the word for ‘small’ in the

8. language
of Rabaul.
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 103
Apart from meri ‘women’ and liklik ‘small’ a number of other
lexical bases are strong candidates for this t y p e of lexical
confla-
tion :
Tolai
atip ‘thatched roof‘
bala ‘belly, bowels’
bulit ‘sap’
rokrok ‘frog’
yat emphasizer
noko ‘midrib of sago’
mom0 ‘to drink’
English
on top
belly
blood
croak croak
Yet
nock, node
more more
T o k Pisin

9. antap ‘on top, roof‘
be1 ‘belly, seat of emotions’
blut, bulut, bulit ‘blood,
sap, glue’
rokrok ‘frog’
yet emphasizer, ‘yet’
nok ‘midrib of sago or
mom0 ‘lots’ (mostly used
feather’
in connection with drink),
excessive drinking
An example involving a New Guinea mainland language and
Eng-
lish is nansei, an exclamation used to attract the attention of
members of the other sex, ‘effeminate man’, which appears to
be
related to both Yakamul nansei ‘sweetheart’ and English nancy.
Roughly analogous combinations of form and meaning can also
be pointed out in cases such as:
Tokai English T o k Pisin
lok ‘to push through’ lock
tak ‘to take’ take tekimwe ‘to take away’
tun ‘to cook, bake’ turn, done tanim ‘to stir food’
dur ‘dirty’ dirty doti ‘dirty’
kap ‘to carry, take’ carry karim ‘to carry, take’
lokim ‘to lock with a key’
Unfortunately, data about the use of Tok Pisin by the indigen-
ous population is very scarce and one can only speculate about

10. the
possible extent of lexical conflation. More than two sources
appear
to have been involved in some instances. A particularly
intriguing
case of lexical conflation is that of sanga ‘pliers, hand of
crayfish,
forked post, slingshot’ which appears to be related to German
Zange ‘pliers’, Malay tiang ‘forked branch’ and Australian
English
104 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
shanghai ‘slingshot’. The meaning shared is that of a
bifurcation
and there appear to be good reasons to assume a triple
etymology
in this case. The last example suggests the importance of two
other
languages in the formation of Tok Pisin: German, the official
language and language of the majority of expatriate households
and missions between 1884 and 1914, and Malay which, in a
pidginized form, was the plantation language and lingua franca
of
the New Guinea mainland (Kaiser Wilhelmsland) before 1900.
The number of lexical items which can be derived equally well
from German or English is quite large, and it can be assumed
that
many of them are the result of conflation, in particular since the
phonological structure of stabilized Tok Pisin provides for the
neu-
tralization of a number of differences in the pronunciation of

11. German and English, such as the treatment of word-final stops.
Consider the following:
Tok
Pisin
ais
anka
as
bet
gaten
hama
mas
rip
sadel
Related
German word
Eis
Anker
Arsch
Bett
Garten
Hammer
Mast
Riff
Sattel
Related
English word
ice
anchor
arse
bed

12. garden
hammer
mast
reef
saddle
Gloss
‘ice’
‘anchor’
‘arse, reason,
origin’
‘bed, shelf‘
‘garden’
‘hammer’
‘mast, flagpole’
‘reef‘
‘saddle’
Numerous additional examples have been pointed out in Stein-
bauer’s dictionary (1969), amounting to more than seventy
items of
shared German and English origin. This is a very significant
number considering that, at the time when both German and
Eng-
lish were available as superstratum languages, the lexicon of
Tok
Pisin comprised between 500 and 1,OOO items, depending on
the
regional origin and degree of acculturation of individual
speakers.
Multiple etymologies involving Malay include the possible re-
inforcement of lexical items which were already established in

13. P. M ~ L H A U S L E R - ETYMOLOGY AND PIDGIN
LANGUAGES 105
German and/or English, including kapok ‘kapok tree’, nanas
‘pine-
apple’ and mango ‘mango’. As both Malay and the Melanesian
languages of New Guinea are members of the same language
family, many instances of cognation can be indicated. Whether
or
not borrowing and mixing has been involved remains unclear.
One
can only appeal to extralinguistic evidence here.3 A Malay
origin
or at least a partial Malay origin has been claimed for a large
number of Tok Pisin lexical items in a paper by Roosman
(1975).
Laycock’s unpublished remarks on the prefinal and final
versions
of this paper form the basis of this discussion.
In a number of cases, the meaning of cognate forms is very
different in Malay and Tok Pisin. Examples are: Malay hormat
‘honour’, which Roosman claims to be one of the sources of
Tok
Pisin amamas ‘to rejoice’; Tok Pisin kalang ‘earring’, which is
said
to be related to Malay kalang ‘circle’. Laycock points out that a
more likely source for amamas is a New Ireland language and
that
kalang in the meaning of ‘earring’ is found in Tolai and related
New Ireland languages.
Even more problematic are cases of conflation of lexical items
originating from different Melanesian languages. What goes for
Malay goes even more for the closely related Melanesian

14. languages
spoken in the areas where Tok Pisin came into being. As
pointed
out by Mosel(l979 : 25): ‘due to the lack of sufficient data from
all
languages which were probably involved in the development of
the
Tok Pisin lexicon, we can only show the possible source
[email protected]
and exclude others, but we cannot definitely state that any Tok
Pisin word is exclusively borrowed from a specific Patpatar -
Tolai
language’. Thus, the mere fact that a word can be traced back to
Tolai is no guarantee that it has actually borrowed from T01ai.~
A
few examples include :
Indigenous
Tok Pisin Gloss Languages Gloss
atap roof, thatch Tolai : etep kunai grass
buai betelnut Tolai : buai betelnut
Mioko : atip thatch
Label : buai
Lamassa : buai
106 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
Pala : buei
Mioko : bue
Molot : bua

15. bulit’ sap, glue Tolai : bulit sap of certain
trees, glue,
blood
Mioko : bulit
Molot : bulit
The presence of lexical items with multiple etymologies thus
poses a number of problems. They are related to the fact that
pidgins are the result of, or accompanying, the gradual
accultura-
tion of a group of speakers. To be more precise :
There are significant cultural and linguistic differences between
the
groups in contact which are only partially bridged in the initial
phases.
Pidgins at the beginning of their life are rather crude makeshift
tongues.
Many writers (e.g. Silverstein 1972) have remarked on the
differences in
grammar and lexicon to be found within a group of pidgin users.
Multi-
ple etymologies can reflect this cultural and linguistic gap. It
cannot be
assumed that speakers and hearers share a lexical item in the
same way
that native speakers in a well-defined speech community do.
Thus, with
a number of words, it is impossible to determine what is the
central and
what the derived ‘metaphorical’ meaning unless reference is
made to the
speaker. Tok Pisin salat is related to both German Snlat
‘lettuce’ and
Melanesian salat ‘stinging nettle’. For a German missionary, to

16. use this
word to mean ‘stinging nettle’ would be a metaphor. For a
Papua New
Guinean the situation would be exactly the reverse.
Schuchardt (1889) makes some relevant remarks on this point:
he
reports that the indigenes of the Duke-of-York Islands have the
words kinkenau ‘to steal’ and tillewat ‘to tie up’ and believe
that
they are of English origin. English speakers, on the other hand,
regard the item tobi ‘to wash from English soap as of Duke-of-
York origin.
So far, we have looked at lexical encounters in the context of
pidgin formation. However, similar chance encounters continue
throughout the development of a pidgin, giving rise to numerous
examples of folk-etymologies. An interesting example involves
the
reinterpretation of a number of English morphemes as Tok Pisin
as
‘origin, foundation’ (from English arse):
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 107
English Tok Pisin Gloss
archbishop as-bishop ‘authoritative bishop’
Ash Wednesday as-trinde ‘important Wednesday’
yesterday as-tete ‘origin of today’
3. DISCONTINUOUS DEVELOPMENIS IN PIDGINS AND
CREOLES

17. When finding the same or a very similar lexical form at an
earlier
and a later stage of a language, linguists working on normal
languages are inclined to assume that :
(i) there has been continuity of transmission;
(ii) therefore the earlier form can be regarded as the etymology
of
the later form (unless of course still earlier forms can be
documented).
Thus, for the English item bee one justified in postulating.
[bi :] [bi :]
‘bee’ ‘bee’ percentage of d using the item speech community
1885 time axis 1975
However, such a conclusion could be quite wrong for a pidgin.
For
the corresponding item and corresponding time span in Tok
Pisin,
for instance, we get:
[bi :]
[binen]
1885 1910 1950 1975
This can be interpreted as follows: between 1885 and 1910 a
declin-
ing number of Tok Pisin users either knew or used the item bi
‘bee’.
However, talking about bees and beekeeping was a rather
marginal

18. domain of the language. By 1910 the original item and the
group of
108 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
users had almost disappeared. A different group (probably
islanders
employed in German households) again had the need to talk
about
bees. The borrowed the item binen from German. Again, by
1950
this group had disappeared without passing on binen to the next
generation. After 1950 a new generation again borrowed bi from
English.
Thus, the word for ‘bee’, along with many other forms referr ing
to concepts marginal to the contact culture, was borrowed at
differ-
ent points in the development of the language only to be subse-
quently lost again. This situation is typical of pidgin languages
which are transmitted between adults and hence lack the
linguistic
continuity of languages learnt by children from their parents or
peergroup (cf. Hockett, 1950). We are thus dealing with a
continu-
ous turnover and recycling of much of the lexicon of a pidgin.
Simply tracing a particular item to the earliest available source
will
fail to do justice to the nature of such a language.
Discontinuity thus appears to be one area where pidgins differ
qualitatively from other languages. It should be noted, however,
that Creoles are no more or less continuous than other

19. languages.
4. THE ‘ETYMOLOGICAL FALLACY’ FALLACY
Lyons (1977 : 244) has characterized the ‘etymological fallacy’
as
follows:
the common belief that the meaning of words can be determined
by investigating their origins.
This view contrasts with the widely accepted one that ‘the
etymol-
ogy of a lexeme is, in principle, synchronically irrelevant’
(ibid.).
There are a number of reasons why this is not necessarily so in
a
pidgin such as Tok Pisin. A first counterargument is that l ike
many
other pidgins and Creoles, Tok Pisin has been in contact with its
original superstratum lexifier language for most of its devel -
opment.6 This contact has been reinforced in the more recent
past
by the introduction of English schooling to a large number of
Papua New Guineans. As a result we are faced with a
continuous
restructuring of semantic and phonological information to bring
Tok Pisin lexical items closer to their putative or genuine
etymo-
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 109

20. logical source. Compare the developments in the semantic area
for
the following items:
Early stabilized Expanded rural
Tok Pisin Tok Pisin
harim harim
smelim
banis banis
banis
peles ples
Sande Sande
wik
peim peim
baiim
Urban
Tok Pisin
hirim
smelim
pilim
listenim
banis
fenis
ples
viles
Sande
wik
holide
peim
baiim

21. spentim
Gloss
to hear
to smell
to feel
to listen to
bandage
fence
place
village
Sunday
week
holiday
to Pay
to buy
to spend
These examples illustrate the operation of language-external
rather than language-internal pressures affecting the narrowing
and
specialization of meaning. In at least three cases, banis, peles
and
peim, it does not seem legitimate to trace the urban form back
directly to early stabilized Tok Pisin. Whereas present-day ples
and
uiles are clearly related to English place and uillage, it would
seem
absurd to trace back uiles to early stabilized Tok Pisin peles, in
spite of the fact that this item is a syncretis m of these two
English
words. I am not even sure whether present day ples can be
traced
back to earlier peles because the semantic narrowing
experienced
by this item appears to be due to renewed contact with English.

22. Discontinuity of transmission is one of the factors accelerating
such restructuring. Rather than learning the stable meanings of
established Tok Pisin items from members of the older
generation,
they rely heavily on the meanings of related English words
learnt at
school.
English etymologies are relevant in yet another area, that of
items with pejorative meaning. The following statement is
perfectly
110 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
reasonable when applied to the stable Tok Pisin spoken in
remote
rural areas:
Why then, I wonder, do speakers of English describe Pidgin as
being full
of insulting words, though they must be aware of the fact that
these
words which bear formal resemblance to insulting words in
English,
have perfectly harmless meanings in Pidgin?
(Wurm, 1967 : 9)
Educated urban Papua New Guineas are now found either to
avoid items which resemble English expletives such as bagarap
‘ruined’ or bulsitim ‘to deceive’, or to use them in the full
awareness
of the connotations they have in Tok Pisin’s lexifer language:

23. Shifts of meaning occasionally take place under the influence of
English,
especially in response to ridicule or disapproval such as that
expressed
by speakers of English toward Neo-Melanesian words or
meanings
which diverge from those of English.
(Hall, 1956)
The number of lexical items thus affected is significant, some
important examples being:
Interpretation in Interpretation in
Lexical item rural T o k Pisin urban T o k Pisin
rabis poor, destitute rubbish, worthless
baksait back backside
pisop to depart quickly to piss off
sarap to be silent, quiet to shut up
In these and similar instances, the continued presence of the
lexifier
language promotes a special type of interlingual word-taboo (cf.
Haas, 1964).
Finally, we can observe, in the history of Tok Pisin, a gradual
change from independent word-formation types to borrowed
ones.
Thus, a ‘raincoat’ in classical Tok Pisin is kot ren, whereas in
many
present-day varieties it is renkot. My feeling is that only the
latter
item should be assigned an English etymology. It should not be
related directly to earlier kot ren, nor should this item, for
reasons
to be outlined, be directly related to English raincoat.

24. One may conclude that the presence of a lexifier language not
only causes post-creole continua at the syntactic level but also
affects continuous lexical restructuring.
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 11 1
5. THE TREATMENT OF LEXICAL CREATIVITY
AND LEXICALLY COMPLEX ITEMS
In direct opposition to the dependency relationship of pidgins
on
outside lexifier languages is lexical creativity encountered in
some,
but by no means all, pidgins and creoles.’ There is considerable
uncertainty as to whether words derived from, or composed of,
lexical bases related to English words should be regarded as of
English origin. Thus, do klinpaia ‘fire which cleans =
purgatory’ or
susoksman ‘shoe and socks man = white collar worker’ qualify
as
items of English origin or not? The common practice in the past
has been to give an affrrmative answer. However, this may just
be
one of the many manifestations of the view that pidgins d o not
have a life of their own but are parasitic upon either a
substratum
or superstratum language.
Thus we find in Mihalic’s dictionary (1972) entries such as:
sakim (E. sack him) to sack s.th., to bag, to put in a bag
pulsen (E. pull chain) a zipper, a hookless fastener

25. pairap (E. fire up) to explode
Similarly Steinbauer (1969) opts for an English origin of
druiwuru
‘low tide, ebb’ and dripman ‘pilgrim, wanderer’. This practice
is
widespread in lexicographical studies of other pidgins and
creoles.
Thus, Bollie (1980 : 71) includes among the 96.7% words of
French
origin in Seychelles ‘creole neologisms formed out of French
lexical
material’ (author’s translation). This practice raises a number of
problems, however, including the following:
(i) it blurs the distinction between clearly borrowed lexical
items
such as renkot ‘raincoat’, calques from English such as manki
spana ‘monkey wrench’, and internal word formation as mani -
fested in manki masta ‘indigenous man in European domestic
employment’;
(ii) it ignores the possibility that compounds may have been
bor-
rowed not directly from English but via other languages. An
example is lukbuk (E. look book?) which is all likelihood was
borrowed from Tolai lukbuk ‘to read’;
112 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
(iii) it does not deal adequately with calques from local
languages
and other instances of lexical conflation, such as the ones men-

26. tioned in section 2.
The most important objection, however, is that the
independence
and vigorous creativity of a language such as Tok Pisin is
simply
ignored. Consequently, in my revision of Mihalic’s dictionary I
have opted for giving an English origin of a complex word only
where direct borrowing is likely, thereby acknowledging the im-
portant role of independent developments in this language.
Such structured, complex lexical items can then be arranged in
the order of their chronological and geographical appearance. If
the distinction between borrowed and language-internal material
is
maintained, such questions as the development of lexical
creativity,
the role of language universals in the derivational lexicon and
con-
straints on borrowing derivational morphology can be meaning-
fully asked. Some of these questions are discussed in
Muhlhausler
(forthcoming).
It can easily be seen that etymologizing for a language such as
Tok Pisin raises, and in some cases provides the answers to, a
number of theoretical questions of historical linguistics. Lack of
time prevents me from going into more detail here. Instead, I
will
devote the remainder of this paper to a number of down-to-earth
practical problems of etymologizing for pidgins and Creoles.
6. PRACTICAL ETYMOLOGIZING FOR TOK PISIN
Some of the problems facing those engaged in etymologizing
for a

27. pidgin, creole or other colonial language have been discussed
by
Dillard (1 970), though researchers working on more ‘normal’
languages will no doubt have had similar experiences. My rather
anecdotal remarks in this section are a response to my discovery
that a large number of etymologies given in available
dictionaries
of Tok Pisin remain highly unsatisfactory. Some of the probable
causes of this state of affairs will be illustrated now.
6.1 Diachronic Purism
The term ‘diachronic purism’ was coined by Valkhoff (1966 : 5)
in
his discussion of theories as to the origin of Afrikaans of the
type
P. MUHLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 113
that ignored linguistic influences other than Dutch and Dutch
dia-
lects. Whereas in the case of Afrikaans diachronic purism
involves
the elimination of non-white influence, in the case of Tok Pisin
it is
found mainly in connection with taboo origins. It refers to con-
scious or unconscious attempts on the part of missionary
lexicogra-
phers to find an innocuous source for Tok Pisin words derived
from English four-letter words or, failing this, simply to
classify
such words as of unknown origin.
A good example of diachronic purism in etymologizing is the

28. derivation of bagarap ‘to be ruined, tired’ from English
‘bankrupt’,
as is done in the Worterbuch mit Redewendungen (around
1935), or
from a non-existant ‘beggared up’, as is done by Schebesta and
Meiser (1949, who comment on their etymology: ‘In English to
beggar is transitive but here the effect is taken’. The correct
deri-
vation from English ‘to bugger up’ does not appear before
Mihalic
(1957). Similar purified etymologies are found for other lexical
items. Compare:
English
Lexicalitem Gloss etymon W M R S M M 7 1 ST69
? kan ‘female cunt -
genitals’
E kok ‘male cock cock - -
genitals’
sit ‘ashes, shit ? not E E
faeces’ listed
- -
( W M R = Worterbuch mit Redewendungen,
SM = Schebesta & Meiser, 1945,
M 7 1 = Mihalic, 1971,
ST69 = Steinbauer, 1969)
Such etymologizing can have side effects, in particular when
used
as the basis of a quasi-etymological writing system. Thus, in

29. devis-
ing a standard spelling system for Tok Pisin the Alexishafen au-
thors of the Worterbuch mit Redewendungen suggest that it
should
closely follow English pronunciation. However, whilst they
spell
114 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
Tok Pisin [han] ‘hand’ as hand they do not restore the final con-
sonant in [kan] ‘female genitals’ as its English origin was not
ac-
knowledged.
6.2 The ‘ifit doesn’t sound English it must be Tolai’principle
There is a very pronounced tendency in all dictionaries of Tok
Pisin to date to trace back most Oceanic, and often most non-
English sounding, lexical material to Tolai, its main substratum
language. This practice has its roots in the neglect of the
complex
socio-historical factors underlying the development of Tok
Pisin.
Until it had been demonstrated that close links existed between
Samoa and New Guinea (Muhlhausler, 1978), for instance, most
words of Samoan origin were given a Tolai etymology.
As pointed out by Mosel (1979 : 23 ff.) a number of lexical
items
commonly listed as of Tolai or Gazelle origin cannot possibly
come
from this source. Two important types of words are:
(i) words which contain the sound [s] which is not found in
Tolai

30. or the Duke-of-York language. This excludes the items b a l m
‘bird’ and melisa ‘barracuda’ listed in Mihalic (1971) and
mosong ‘fluff and susu ‘breasts’, for which Tolai has been
given
as the source by other lexicographers.
(ii) words containing no prenasalization before voiced stops.
Thus,
neither rabun ‘ridge of house’ nor abus ‘animal’ can be of Ga-
zelle origin (as assumed by Mihalic, 1971); the latter item ad-
ditionally contains the sound [s].
The last-mentioned item is not only a phonetologically unlikely
transfer from Tolai; there simply is no Tolai item which could
be
regarded as a possible cognate of abus.
The solution to this, and possibly quite a few other unsolved
etymological problems in Tok Pisin, lies in the reliance on
chance
discoveries, flashes of imagination and so forth, helped along
by
recent developments in natural phonology and semantics. My
pro-
posed source for abus is English animals. The transition from
ani-
mals to abus can be accounted for by a number of highly natural
processes:
(i) I becomes u because of their close accoustic similarity. This
yields animus;
P. MWLHAUSLER - ETYMOLOGY AND PIDGIN

31. LANGUAGES 115
(ii) the least prominent syllable is lost, to yield amus;’
(iii) the more highly marked nasal is replaced with a
homorganic
I have to confess that this possibility only occurred to me when
my
then 21-month-old daughter began to refer to animals as abus.
Another item of alleged Tolai origin is pui ‘naked’. Again, no
plausible cognate has been found. The solution to this problem
was
suggested to me by Dr. John ZGraggen, a Swiss missionary
linguist. His proposal was German pfui ‘exclamation of disgust’
uttered by German missionaries when faced with local nudity.
stop, yielding abus.
6.3 Ethnocentricity
Nothing is more difficult for linguists than avoiding the
projection
of linguistic and metalinguistic knowledge from their own
culture
to foreign languages, in particular exotic ones, under
investigation.
This tendency is particularly strong when the foreign language
is
known by a name such as Pidgin English or Creole French.
Thus, next to cases where English words are treated as of Tolai
origin one finds examples of attempts to trace exotic words to
familiar English ones. For instance, instead of recognizing Tolai
pekapeke ‘to defecate’ as the most likely source, Aufinger
(1949 : 118) proposes the following explanation:

32. The word ‘bekbek’ for defecation is probably derived from the
habit of
the natives along the beach to ease themselves along the water-
line, thus
forming a line of many backs, expressed by the reduplication
‘bekbeVg
Mihalic (1973) derives kanaka ‘indigene’, a word widely known
in
the pidgins of the Pacific from English cane hacker instead of
Ocea-
nic kanaka ‘man, boy’. It is true that a number of indigenes
were
employed in the sugar industry, but it must be noted that the
word
existed long before the sugar industry and is used in areas
where
neither is sugar grown nor did local men go to work on sugar
plantations.
Finally, the origin of mumut ‘large bushrat’, from Duke-of-York
mumut, is given as English marmot in a number of dictionaries.
One may feel inclined to laugh at such explanations, but it may
be better to remember what was said about lexical encounters
earlier in this paper. In language contact situations such as are
116 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
characteristic of the formation period of pidgins, many of the
par-
ticipants do not know which words belong to which language. In
the beginning, many words, and I am prepared to go so far as to

33. say the majority, belong to more than one of the languages in
contact. As a pidgin develops, changing social patterns, in
particu-
lar the growing dominance of the expatriate colonizers, can
obliter-
ate the original syncretisms and bring about a closer association
with what Western linguists have come to call the ‘lexifier
language’.
7. CONCLUSIONS
The principal problem of etymologizing for pidgins and Creoles
would seem to be the result of (i) their greatly accelerated
linguistic
development and (ii) the operation of catastrophic rather than
continuous forces. This means that whereas many of the lexical
developments of these languages are perfectly comparable to
devel-
opments in ordinary languages and at best quantitatively
different,
there remain some areas where pidgins and Creoles may differ
more
radically and where traditional methods of etymologizing may
not
apply. Instead, more refined methods, taking into account
different
types of transmission, will have to be developed.
In summarizing the findings of my paper I would like to pro-
pose :
(i) While it is possible to trace the etymologies of lexical items
in
pidgins and Creoles, there are some special problems with these
languages. In order to overcome them, close attention has to
be paid to the socio-historical context in which they develop.

34. (ii) It appears that, at some point of development, a large
propor-
tion of pidgin and creole lexical items can be assigned jointly
to more than one source language. The traditional notion of
‘lexifier language’ stands in need of revision.
(iii) Etymologies can change as a result of discontinuities in
trans-
mission and prolonged contact with lexically related
languages.
(iv) Borrowing tends to conceal its traces. Cognates can result
from either shared history or chance encounters.
P. MWLHAUSLER - ETYMOLOGY AND PIDGIN
LANGUAGES 117
Etymologizing for pidgins and Creoles has only just begun. It is
hoped that more pidginists and creolists will turn to
etymological
questions and that their research can give a renewed impetus to
a
branch of linguistics which has been neglected for too long.
Linacre College
Oxford
NOTES
1. Similar lexical items are also found in the other languages of
the Duke-of-York
New Britain and New Ireland area, thus further reinforcing the
choice of be1 rather

35. than equally possible pidgin words such as tummy or tumtum.
2. I have on record the following variants: kiap, kapen, kiapen,
kiapi, kapten.
[P.M.]
3. The most pertinent questions being: Was such a lexical item
likely to have
been used in the plantation context? and Does it refer to an
object introduced by
Malay bird-of-paradise hunters in the area?
4. The problem is similar to that of identifying Germanicisms in
Dutch.
5. There is also a German etymology (Blut = blood) for this
item.
6. Even in times of German control over New Guinea, a
significant number of
English speakers remained in the colony and may have
continued to serve as a
lexical model.
7. For reasons unknown, the lexical creativity of Tok Pisin is
greater even than
that of many Creoles. A full account is given by Muhlhiiusler
(1979).
8. This process is also found in other polysyllabic words used
by my daughter,
such as elan for elephant and pekan for pelican.
9. Note that in some secret varieties of Tok Pisin the
replacement for pekpek ‘to
defecate’ is si ‘to go to the seashore’.

36. REFERENCES
AUFINGER, A., 1948 & 1949. ‘Secret Languages of the small
islands near Madang’
BAILEY, C. J. N., 1980. ‘The Role of Language Development
in a Theory of
BAIWN, G., 1944. ‘Pidgin English and Cross-Cultural
Communication’, Trans-
BOLLBE, A., 1980. ‘Zum Projekt eines Dictionnaire
Etymologique de Creole’, Roman-
CASIDY, F. G., 1966. ‘Multiple Etymologies in Jamaican
Creole’, American Speech
DILLARD, J. L., 1970. ‘Principles in the History of American
English - Paradox,
Virginity and Cafeteria’, The Florida FL Reporter Spring/Fall
1970.
EDWARDS, J., 1974. ‘African influences on the English of San
Andres Island, Colum-
bia’, in De Camp & Hancock (eds.), Pidgins and Creoles,
Georgetown UP, 1-26.
South Pacific Vol. 3,4,90-95; Vol. 3,5,113-120.
Language’, Papiere zu tinguistik 22, 33-46.
actions of the New York Academy ofsciences 2, 137-141.
ica Europaea et Americana, Bonn: Bouvier Verlag, 68-76.
32,49-53.

37. 118 TRANSACTIONS OF THE PHILOLOGICAL SOCIETY
1982
HAAS, M. R., 1964. ‘Interlingual Word Taboos’, in Hymes
(1964), 4 8 9 4 9 4 .
HALL, R. A., Jr., 1956. ‘Innovations in Melanesian Pidgin (Neo
Melanesian)’,
HALL, R. A., Jr., 1966. Pidgin and Creole Languages, Ithaca,
Cornell University
HYMES, D. (ed.), 1964. Language in Culture and Society,
Harper & Rown & John
HYMES, D. (ed.), 1971. Pidginization and Creolization of
Languages, Cambridge Uni-
LAYCOCK, D. C., 1970. Materials in N e w Guinea Pidgin
Pacijic Linguistics, Series D.
LE PAGE, R. B., 1974. ‘Processes of Pidginization and
Creolization’, Yurk Papers in
LYONS, J., 1977. Semantics, 2 vols., Cambridge University
Press, London.
MALKIEL, Y., 1975. ‘Etymology and Modern Linguistics’,
Lingua 36,101-120.
MCELHANON, K. A. (ed.), 1975. T o k Pisin: Go W e ? ,
Linguistic Society o f Papua
MIHALIC, F., 1971. T h e Jacaranda Dictionary and Grammar
of Melanesian Pidgin,
MIHALIC, F., 1973. ‘Stori bilong Tok Pisin (4)’, Wantok

38. Niuspepa 24 August, 8.
MOSEL, U., 1979. Tolai and T o k Pisin, Pacific Linguistics B-
73, Canberra.
MUHLHAUSLER, P., 1978. ‘Samoan Plantation Pidgin English
and the Origin of New
Guinea Pidgin’, Papers in Pidgin and Creole Linguistics No. 1,
Pacific Linguistics
A-54,67-120, Canberra.
MUHLHAUSLER, P., 1979. Growth and Structure of the
Lexicon of N e w Guinea Pidgin,
Pacijic Linguistics C-52, Canberra.
MUHLHAUSLER, P., forthcoming. ‘The Development of
Derivational Morphology in
Tok Pisin’, Folia Linguistica.
NEVERMANN, H., 1929. ‘Das Melanesische Pidjin-Englisch‘,
Englische Studien 63,
RDOSMAN, R. S., 1975. ‘The Malay Element in Melanesian
Pidgin’, in McElhanon
SCHEBESTA, J. and MEISER, L., 1945. Dictionary of ‘Bisinis
English’ (Pidgin English),
SCHUCHARDT, H., 1889. ‘Beitrage zur Kenntnis des
englischen Kreolisch: 11.
SILVERSTEIN, M., 1972. ‘Chinook Jargon: Language Contact
and the Problem of
STEINBAUER, F., 1969. Concise Dictionary of New Guinea
Pidgin ( N e o Melanesian)’,

39. VALKHOFF, M. F., 1966. Studies in Portuguese and Creole,
Witwatersrand University
WHINNOM, K., 1971. ‘Linguistic Hybridization and the
‘Special Case’ of Pidgins and
W~RTERBUCH, Undated (probably 1935). Wiirterbuch mit
Redewendungen, M S Alex-
WURM, S. A., 1967. ‘Papua New Guinea Nationhood: The
Problem of a National
Oceania 26,91-109.
Press.
Weatherhill Inc., New York & Tokyo.
versity Press, London.
no. 5, Canberra.
Linguistics 4,41-69.
New Guinea, Port Moresby.
Jacaranda Press, Brisbane.
252-258.
(1979,229-234.
MS Alexishafen PNG.
Melaneso-Englishes’, Englische Studien 13, 158-162.

40. Multi-Level Generative Systems’, Language 48/2, 3 7 3 a ;
48/3,596-625.
Kristen Press, Madang PNG.
Press, Johannesburg.
Creoles’, in Hymes (1971), 91-116.
ishafen PNG.
Language’, Journal of the Papua N e w Guinea Society 1,1, 1-
13.