Transcript A Sociolinguistics of Loan Words Classification in Zulu

A Sociolinguistics of Loan Words
Classification in Zulu Language
M.N. Ngcobo
Department of Linguistics
Outline
Introduction
•
•
•
•
•
•
•
Loan words:- words which are formed through borrowing and this is evident where
languages co-exist side by side in the same community.
Most new words enter Zulu language by borrowing roots from another language
such as English and Afrikaans.
These words are borrowed and adapted to the grammatical system of Zulu
language. E.g., the English word ‘radio’ will be borrowed and adapted as ‘i-radio’ in
Zulu language.
However, most Zulu native speakers will still recognize this loan word as a foreign
word.
But some loan words have been assimilated to Zulu language to such that native
speakers are surprised that they are borrowed words (Schnoebelen 2005). E.g.
‘ibhulukwe’ which is borrowed from Afrikaans ‘broek’.
The linguistic phenomenon referred to the first example is called lexicalization
while the latter example can be considered as institutionalization (cf. Al-Azzam et
al. 2008).
Lexicalization deals with the creation of new words (Lipka 1990:95) while
institutionalization refers to the coming to general use of words in society and
being listed in dictionaries (Bauer 1988:67).
Focus of the study
• In this study I address the issue of classifying loan words in Zulu
language, a phenomenon that is evident in the formation of new
lexical items through borrowing from English and Afrikaans sources.
• Taken from stems words from these lending languages are
borrowed and lexicalized to coin new words which are later
institutionalized with the everyday use.
• I concentrate on nominal morphological restructuring of Zulu loan
words in addition to the encoding system they have gone through
at the semantic level from a sociolinguistics perspective.
• The underlying assumption of this study is that various speakers of
Zulu language classify loan words in different noun classes based on
their sociolinguistic background.
Zulu language and Noun Classification
South-eastern
Bantu Zone
Nguni group
Sotho group
Venda group
Tsonga group
IsiZulu
Southern
Sotho
Ronga
IsiXhosa
Sesotho sa
Lebowa
Tonga
IsiNdebele
Setswana
Tswa
SiSwati
Classification
• As part of Bantu language family Zulu language
uses numbered classification system for the noun
class prefixes and this has become the standard
format for listing the noun class prefix system of
any specific Bantu language since Meinhof
(1906).
• According to Hendrikse and Poulos (1994:230):
Traditionally the noun universe in Bantu is divided into 23 classes.
Underlying the traditional view of the Bantu class system is a
typical view of categories. In terms of this view, every member of
a noun should unequivocally satisfy certain specifiable linguistic
criteria.
Zulu noun classes
The Zulu noun class system consist at least 16 of the 23 Bantu classes as shown in the table below.
1
umu-/umumuntu/umlungu
Person/European
Human beings
1a
uubaba
Father
Proper names, kinship terms, personifications
2
aba-/abe-/ababantu/abelungu/aboni
People/Europeans/sinners
Regular plural of Class 1
2a
oobaba
fathers
Regular plural of Class 1a
3
umu-/um-/uumuthi/umthombo/unogwaja
Tree/stream/hare
Natural phenomena, body parts, plants, animals
4
imi-/imimithi/imeqo
Trees/wizard spell
Regular plural of Class 3
5
ili-/iilitshe or itshe
Stone
Natural phenomena, animals, body parts, collective nouns, undesirable people, augmentatives, derogatives
Zulu noun classes...
6
ama-/ameamatshe/amehlo
Stones/eyes
Regular plural of Class 5, mass terms and liquids, time references, mannerisms, modes of action
7
isi-/isisisu/isono
Stomach/sin
Body parts, tools, instruments and utensils, animals and insects, languages, diseases, outstanding people, amelioratives, derogatives, diminutives, augmentatives, curtatives (shortness and
stoutness), mannerisms
8
izi-/izizisu/izono
Stomachs/sins
Regular plural of Class 7
9
iN(n,m)inkomo/imvu
Cow/sheep
Animals, people, body parts, tools, instruments and household effects, natural phenomena
10
iziN(n,m)izinkomo/izimvu
Cows/sheep
Regular plural of Class 9
11
ulu-/uuluthi or uthi
Stick
Long, thin entities, languages, body parts, natural phenomena, implements, utensils and other artefacts, augmentatives, derogatives
14
ubu-/ub-/uubuso/uboya/utshani
Face/hair/grass
Abstracts, collectives, locations, infinitives
Zulu noun classes
15
uku-/uk-/ukwukudla/ukona/ukwakha
Food/to sin/to build
Infinitives
16
phaphandle
outside
Locations
Zulu noun classes
• Classes 12 and 13 are not represented in Zulu noun classification system
as these indicate dimunitive constructions which are only marked by a
dimunitive suffix –ana rather than a prefix in Zulu language.
• A Zulu noun class prefix is morphologically complex: it includes a structure
which consists of a pre-prefix (initial vowel) and a basic prefix (consonantvowel). E.g. umuntu / umu(P)-ntu(STEM) / u(IV)-mu(BP)-ntu(STEM)
• Zulu singular classes have either vowel [i] or [u] as pre-prefix.
• Semantically, the classes are distinguished according to their semantic
properties of the range of nouns that instantiate them.
• However, Hendrikse and Poulos (1992:202) have noted that there is some
semantic heterogeneity in each class, yet each class seems to have a
common underlying denominator that supersedes the semantic
heterogeneity evident in the meanings of the nouns associated with the
class.
The assumptions about the nominal classification of loan words in Zulu
language
•
•
•
•
•
The noun class:- traditionally has been regarded as an arbitrary morphological
system (cf. Zawada and Ngcobo 2008).
Recently, the focus has shifted to the semantic content.
For example, classes 1a/2a and 1/2 are associated with nouns that refer to human
beings and they are dominated by personal nouns.
The classification of nominals that are derived from loan words follows a similar
semantic pattern.
Canonici (1995) has classified loan words under what he calls sub-classes 3a and
9a. He describes class 3a as a class that contains borrowed nouns of impersonal
objects, ‘mass’, food stuffs which has a prefix u-, for example, ubheseni (basin)
(Canonici 1995:9). He asserts that this class has no plural, but the few nouns which
are occasionally used in the plural, make use of class 2a, for example obheseni
(basins). He argues that this cannot be considered as part of class 1a because
these impersonal nouns make use of the concords of class 3 (Subject Concord u-;
Object Concord -wu-; Absolute pronoun wona (not yena), etc.). Following a similar
pattern Canonici classifies all borrowed nouns that begin with a prefix i- in class 9a.
The assumptions...
• While Canonici has tried to justify the sub-classes for borrowed nouns in
Zulu language, he has failed to explain why some borrowed nouns such as
udokotela (dokter) have been classified in class 1a.
• Class 1a is traditionally a class that contains personal nouns which refer to
professions. Since udokotela refers to a profession it is thus classified in
class 1a.
• There is therefore no reason to classify those nouns that have a prefix u- in
a sub-class if they share similar semantic properties with those in class 3.
Nor we cannot classify borrowed nouns in sub-class 9 if they share similar
properties with class 9.
• Canonici’s perspective has also failed to realize the fact that borrowed
words can be classified in more than just two classes (therefore more than
just two sub-classes had to be invented if we were to follow that route).
One may find loan words also classified in classes 5 (irayisi – rice) and 7
(isitofu – stoof).
Assumptions...
•
•
•
•
•
•
The person who has come closer to the analysis of the classification of these loan
words in Zulu language is Schnoebelen (2005).
Looking at how semantic marking is applied to loan words that are brought into
Zulu and how these words are assigned to classes, Schnoebelen has classified all
personal loan words such as unesi (nurse) in classes 1a/2a.
He also classifies all other loan words such as umasikido (mosquito) in classes
1a/2a.
Schnoebelen has also noticed that some borrowed nouns are classifiable in class 5
and he cites Wilkes (1995) where he asserts that most loan words occur in the iliclass.
He also classifies some loan words in class 9 and he notes that “Canonici wants to
place these nouns in a new subclass, called 9a…”. He quotes Canonici asserting
that this is “a growing new class of borrowed nouns in Zulu…” (Schnoebelen
2005:6).
He is also aware that there are some borrowed nouns that are classified in class 7
such as isitezi (stairs). He thinks that this is due to phonology and misanalysis as he
argues that “if the foreign word’s initial sounds are similar to an existing prefix
those sounds can be heard as the prefix” (Schnoebelen 2005:8).
Theoretical framework
• Recently, the principles of cognitive grammar have become a useful
tool for investigating the semantic structure of noun classes.
• This approach explicitly recognizes the fact that human beings use
linguistic categories to make sense out of the world, and it provides
a cognitively motivated framework for describing associative
relations among the members of a category.
• The principles by which different nouns are grouped together into a
class are similar to those that govern the connection among the
various senses of an individual lexical item.
• These same principles, including metaphor and metonymy, also play
an important role in lexical and grammatical change (Heine and
Claudi 1986).
Theoretical framework
•
•
•
•
•
•
•
•
While this perspective has open up a forum of looking at the grammatical structures in terms of
their function (communicating semantic content), there is still a need to look at the sociolinguistic
function of these structures.
This includes facts about who uses them and about how often and under what circumstances they
use them (Hudson 1986:1).
“It is POSSIBLE” to integrate sociolinguistics facts and structural facts into the same formal structure
– in other words, we can include sociolinguistic facts in grammars (Hudson 1996:4) .
Sociolinguistics is responsible for the study of interactions between structural relations and register.
The assumption that some pattern is used or avoided by all speakers of a language can be
addressed by a sociolinguistic analysis by looking at a social distribution of pattern.
Hudson (1986:5) argues that the present state of affairs in mainstream structural linguistics does
little to encourage sociolinguistics to contribute, and even less to encourage structural linguistics to
look to sociolinguistics.
According to him speech is a kind of action, so knowledge about speech is a kind of knowledge
about action. He looks at language as a mode of this action (speech). He argues that the typical
action has an actor so we must conclude ... that the typical word also has an actor (Hudson
1986:10).
From this perspective I will argue that we need to investigate social constraint in loan word
classification in the Zulu language. This allows us to ask a question: what kind of a person uses this
kind of structure. This is based on an observation that some nouns are restricted to use by certain
types of speaker.
Methodology
•
•
•
•
•
•
•
•
A corpus-linguistics approach: to make observations regarding frequency and sociolinguistic
patterns of Zulu borrowed nouns as they are reflected in a spoken language corpus of actual
language usage.
When we consider the fact that speech is action and that language is mode of this action (speech),
it is therefore important to keep written language out of consideration.
For example, some forms found in speech are not written and therefore not in written language.
I will therefore use a corpus of collected spoken spontaneous speech to see which classes are used
to classify loan words in Zulu language and that which classes of loan words occur most frequently
in normal spoken language.
Once the noun classes have been identified the borrowed nouns will be classified accordingly. This
will help to observe which noun classes classify more borrowed nouns than others.
Thereafter, I will attempt to describe the type of sociolinguistic factors that play a role in the
classification process.
The data for this study are drawn from a spoken language corpus of Zulu of approximately 20 000
tokens.
I also compare this corpus with the sample Xhosa corpus of the same size in order to see how
another Nguni language which is closer to Zulu classify these new words. Both Zulu and Xhosa
spoken language corpora is part of the Unisa Southern African Spoken and Signed Language
Corpora project (SASSLC). This corpus can be considered as representative in terms of a variety of
speakers (in respect of gender, age and educational background) and a variety of registers, settings
and activities. Natural language was video-recorded and then transcribed and stored.
Methodology
•
•
•
•
•
•
•
•
•
•
The first step: collating the transcripts of Zulu spoken language transcribed data.
The next step: generating various word lists with all words in the transcripts by using WordSmith
Tools (both alphabetically and in terms of frequency).
The word lists were used to identify all loan words (nominals) and their respective noun classes
manually since Zulu corpus is not as yet annotated for either parts of speech or for morphological
classes.
All inflected nominals were removed in favour of uninflected nouns.
Finally, Wordsmith Tools was used again to list all frequencies of these borrowed nouns in order to
determine frequencies per noun class.
Therefore, WordSmith Tools helped to quantify the occurrences of the borrowed nouns, and of the
nouns in each of the classes.
I selected and classified borrowed nouns in their respective classes by simply looking at their
prefixes.
To avoid my own subjective view of noun classification, I used another tool called Concordance to
see how these nouns are used in context.
For example, the noun iselulafoni (cellphone) can be classified either in class 7 or class 9
respectively depending on the speaker.
In order to make comparisons I also extracted the spoken language Xhosa corpus of the same size
and I followed the same procedure for comparitive reasons.
Data presentation and analysis
• I applied WordSmith Tools to isolate each token that is found in the
transcribed spoken corpora.
• My task was to extract borrowed words from the list and note their
frequency of use in the transcribed spoken language Zulu corpus.
• The tools gave me access to the ranked frequency of all words in
the sample corpora of Zulu language.
• By using the token lists I was able to obtain a sample of single
words extracted from the transcribed spoken corpus data.
• The lists called List A and List F respectively provided me with
alphabetically arranged words and the frequency of their
occurrence. The Concordance tool was used to see the words in
context.
Results
•
•
•
•
In the transcribed spoken Zulu corpus data of exactly 18954 tokens and 5584 word types I
found about 229 tokens representing borrowed words.
That counted for 77 word types in the total sample corpus.
By contrast, in the transcribed spoken Xhosa corpus of exactly 19238 tokens and 5143 word
types I found 289 tokens representing new words.
This counted for 85 word types in this sample corpus.
Presentation
ZULU
XHOSA
CLASS
Frequency
Types
Frequency
Types
1a/2a
35
9
107
11
1/2
4
3
14
3
3/4
9
7
11
6
5/6
30
20
10
8
7/8
66
16
49
14
9/10
71
17
98
43
11/10
5
3
14
10
3
15
7
3
Results
• The observation here is that there is no significant difference
between Zulu and Xhosa in terms of the number of borrowed
words in this sample.
• Another observation is that while Zulu classify borrowed words in
almost every class, Xhosa only classifies these words in specific
classes.
• In an unrelated study by Ngcobo and Nomdebevana (2010,
forthcoming) it was observed that these borrowed words can be
categorised into the various groups according to how they are
borrowed and adapted in the indigenous languages. While some of
these words are adoptives that have been phonologically
assimilated to the Zulu system, some are coinages that have been
formed from a native perspective of what the borrowed object or
phenomenon might mean such as the word ‘ingculaza’ (HIV/Aids).