Sense Relations in Dholuo: A Lexical Pragmatic Approach

1Introduction

Dholuo is a Western Nilotic language spoken by the Luo people of western Kenya and northern Tanzania. Related languages in the same group include Acholi, Alur, Dinka, Lango and Shilluk. The majority of Dholuo speakers in Kenya are found in Kisumu, Homa Bay, Siaya and Migori counties. Stafford (1967) identifies two mutually intelligible dialects distinguished by vocabulary and pronunciation, which he terms South Nyanza and Trans-Yala. Oduol (1990) identifies the same division as Kisumu–South Nyanza (KSN) and Boro–Ukwala (BU). The Kisumu–South Nyanza dialect is spoken across a wider geographical area and is consequently regarded as the standard form, while Boro–Ukwala is spoken in a smaller region and carries less overt prestige.

All languages allow speakers to describe aspects of what they perceive. In Sapir’s formulation, quoted in Saeed (2003), the particular language we speak conditions our conceptualisation of the world. Saeed (2003) further observes that what a speaker refers to by using any linguistic item depends on context, so that pragmatics occupies a central place in the meaning-making process of any language. Sense, by contrast, concerns the system of relationships holding between linguistic elements themselves, principally between words (Palmer, 1976).

According to Lyons (1968), the sense of a lexical item may be defined not merely as dependent upon, but as identical with, the set of relations holding between the item in question and other items in the same lexical system. Certain vocabulary items fall into lexical systems whose semantic structure is to be described in terms of the sense relations holding among their members (Murphy, 2003).

Relations of this kind may be paradigmatic or syntagmatic (Lyons, 1981). Paradigmatic relations are those into which a linguistic unit enters by being contrasted with, or substitutable for, other similar units in a particular environment; they include synonymy, polysemy and homonymy. Syntagmatic relations are those a unit contracts by virtue of its co-occurrence with other units. A lexical pragmatic approach departs from formal or purely descriptive accounts of semantic relations, in which words are considered only with reference to their denotative meaning, and examines instead the contexts of usage on which interpretation depends. This paper sets out to establish the sense relations obtaining in Dholuo and to show how paradigmatic relations in the language distinguish different kinds of sense relation.

1.1Research objective

The study pursues a single objective: to identify and describe the sense relations obtaining in Dholuo within a lexical pragmatic framework, with particular attention to the contextual factors that condition their interpretation.

The scope of the study is confined to the paradigmatic relations of synonymy, homonymy and polysemy. Antonymy, hyponymy and meronymy, although equally paradigmatic, are not treated here and remain a subject for further work.

2Theoretical framework and related literature

2.1Lexical pragmatics

The study adopts lexical pragmatics as its theoretical framework. The approach was proposed by Blutner (1998) and developed further by Carston (2002) and Wilson (2003), who addressed problems and inadequacies arising in its earlier formulation. Lexical pragmatics moves beyond encoded meaning to analyse meaning in the context in which a word occurs, and it distinguishes two principal pragmatic processes, narrowing and broadening, through which the concept communicated comes to differ from the concept encoded.

Lexical semantics approaches the study of concepts from a linguistic perspective and investigates the properties concepts must have in order to function adequately as word meanings. Blutner (1998) demonstrates that the difference between linguistically encoded semantic structure and conveyed meaning follows from general principles of cooperative behaviour. Carston (2002) argues that an adequate lexical pragmatics requires a formulation of conversational implicature alongside the notion of semantic underspecification in the lexicon, together with an appropriate representation of contextual and encyclopaedic knowledge. Narrowing and broadening are therefore not peripheral adjustments but the ordinary mechanism by which lexical items are interpreted in use, and it is this claim that makes the framework suitable for the present data.

2.2Studies of sense relations

Murphy (2003) examines the paradigmatic semantic relations holding between words, including synonymy, antonymy and hyponymy, and their relevance to the mental organisation of the vocabulary. Sweetser (1991) analyses the multiple meanings of English modals, conjunctions, conditionals and perceptual verbs, arguing that meaning extends beyond human cognitive structure to encompass the metaphorical and cultural aspects of that structure. The present study draws on both in confining its attention to the open word classes.

Atoh (2002) examines the semantic properties of Dholuo nouns with a view to investigating their relationships within given sentences. The central claim of that study is that the choice of nouns used in a given discourse event by native speakers is informed by the distinctive components of the words and by their semantic relationships within and across semantic fields. The present study extends this line of inquiry beyond the noun to the open class generally, and shifts the analytic frame from the semantic field to lexical pragmatics.

Ndlovu (2001) investigates sense relations in Isichazamazwi SesiNdebele, focusing on hyponymy, meronymy and synonymy. Ndlovu argues that hyponyms and meronyms facilitate the formulation of more user-friendly and accessible dictionary definitions, whereas definitions by synonym reduce accessibility. The study is relevant here for its demonstration that sense relations carry consequences for lexicography in African languages.

Mwebia (2006) explores sense relations in Kimeru using lexical pragmatic theory, considering the differences and similarities in sense relations involving corresponding lexical items in English and Kimeru. Mugure (2009) discusses sense relations in Gikuyu, treating synonymy, homonymy and polysemy together with semantic narrowing and broadening, and concludes that lexical pragmatic theory works adequately in analysing homonymy and polysemy under the scope of narrowing or broadening but cannot sufficiently handle synonymy. The present study tests that conclusion against Dholuo data.

2.3Research gap

Dholuo has received sustained attention in phonology, morphophonemics and syntax (Okombo, 1982; Omondi, 1982; Oduor, 2002) and in the semantics of the noun (Atoh, 2002), and lexical pragmatic analyses of sense relations exist for Kimeru (Mwebia, 2006) and Gikuyu (Mugure, 2009). No comparable lexical pragmatic account of sense relations in Dholuo has been undertaken. The present study addresses that gap, and in doing so provides a point of comparison for the Bantu-language studies cited, since Dholuo is Nilotic.

3Methodology

3.1Study area and design

The study was carried out in Siaya, Kisumu and Migori counties. These regions were selected because they contain the majority of Dholuo native speakers (Oduol, 1990). To ensure competence in the variety under study, informants were not drawn from border regions, so as to avoid interference from neighbouring languages.

The study adopted a descriptive research design. The approach is appropriate to qualitative data of this kind because it provides systematic, descriptive and explanatory accounts of categories or classes of lexical items.

3.2Sampling and participants

A stratified random sampling technique was used to select 200 informants across the three study areas. Informants were stratified by gender, age and level of education so as to obtain proportional representation of each stratum, and were selected with a view to generalising from the sample to the wider speech community. The purpose of stratification was to ensure that the characteristics of speakers most likely to condition lexical choice were represented in the data.

3.3Data collection

A structured questionnaire was used to elicit lexical items exhibiting the sense relations of synonymy, homonymy and polysemy. Respondents who could not read or write were prompted orally by the researcher, and their responses recorded in the same instrument. Each elicited item was matched with an English gloss and classified under the relevant sense relation. Where the information obtained from respondents was insufficient to settle a classification, the researcher supplemented the data by drawing on native-speaker competence.

A target of 500 lexical items was set at the outset, drawn from the open word classes, of which 300 satisfied the criteria for classification and were retained for analysis.

3.4Data analysis

The elicited lexemes were classified and coded according to word class and then examined to determine the nature and patterns of the semantic relationships holding among them. Classification followed the definitional criteria set out in Section II, and each item was then considered in the contexts of use supplied by informants in order to establish the narrowing or broadening operations involved in its interpretation.

4Findings and discussion

The analysis identified three sense relations operating in Dholuo: synonymy, homonymy and polysemy. These relations demonstrate that lexical items do not function as isolated units but are interconnected through semantic networks. Within synonymy, the data further display systematic variation along five dimensions: formality, dialect, borrowing, emotive meaning and collocation. Each is treated in turn below.

4.1Synonymy

Cruse (2000) observes that if synonymy is interpreted simply as sameness of meaning it appears a rather uninteresting relation, but that if synonyms are taken to be words whose semantic similarities are more salient than their differences, a substantial area of interest opens up. Ullmann (1970) notes that very few words are completely synonymous in the sense of being interchangeable in any context without the slightest alteration in objective meaning, and Palmer (1981) states the position more strongly: there are no real synonyms, and no two words have exactly the same meaning. The Dholuo data support the weaker formulation. Items that are substitutable in some contexts prove non-substitutable in others, and the conditioning factors are pragmatic rather than denotational.

4.1.1Formality

Dholuo possesses sets of lexical items sharing a denotation but differing in formality. The examples in Table 1 all denote to die.

Formal register	Informal register
ng’ielo orengo	tho
nindo e bat	lengó lep
yweyo	gwe tao

Note. Source: field data. The degree of formality determines the context in which each synonym may be used.

A related but distinct set consists of euphemisms for dying, each of which carries an independent literal sense elsewhere in the language. A euphemism is a word or expression used to speak of unpleasant or embarrassing subjects without naming the things themselves. These items are set out in Table 2.

Lexeme	Literal sense
yweyo	rest
nindo	sleep
luongo	call
weyo	leave
kao chuny	take a soul

Note. Source: field data.

The euphemistic use of these items is a clear case of pragmatic narrowing in Blutner’s (1998) sense: the encoded concept rest or sleep is narrowed in the appropriate context to a specific concept of dying that the encoded form does not itself express. The overlap between Tables 1 and 2, where yweyo and nindo appear in both, is not an inconsistency but evidence that the formality and euphemism dimensions interact: the most formal expressions in this field are precisely the euphemistic ones.

4.1.2Dialectal variation

Palmer (1981) observes that one member of a synonymous pair may belong to a different dialect from the other, with one variety treated as standard and the other as regional. In Dholuo, although each dialect has its own lexeme, speakers of both dialects may use either item in communication. In certain cases, however, the lexemes are not interchangeable because they denote distinct concepts. Examples are given in Table 3.

Kisumu–South Nyanza	Boro–Ukwala	Gloss
nyathi	mihia	baby
puodho	ndalo	garden
kor ot	sisi	wall
pidho	gago	feed

Note. Source: field data.

4.1.3Borrowing

Borrowing has introduced synonyms into the Dholuo lexicon. Borrowing refers to linguistic items being taken over by one language or dialect from another (Crystal, 1997). The borrowed item and the inherited item typically coexist, as shown in Table 4.

Borrowed item	Inherited Dholuo item	Gloss
mtoka	nyamburko	motor car
osiptal	odthieth	hospital
pesa	omenda	money

Note. Source: field data. The borrowed items derive from Kiswahili and, ultimately in the case of mtoka and osiptal, from English.

These pairs come closest in the data to absolute synonymy, since the two members may share denotation, register and collocational range, and either may be selected without altering the proposition expressed.

4.1.4Emotive meaning

Synonyms in Dholuo also differ in the attitudes associated with one member of a pair, so that a lexeme may convey a positive or a negative evaluation of the same referent. Table 5 presents these pairs.

Positive	Negative	Gloss
miyo	dhako	woman
pek	yach	pregnant
tegno	opong’	mature

Note. Source: field data.

Evaluative differentiation of this kind cannot be captured in a denotational account, since both members pick out the same referent. It is accommodated naturally in a lexical pragmatic framework, in which the evaluative component is treated as part of the concept constructed in context rather than as part of the encoded lexical entry.

4.1.5Collocation

Palmer (1981) notes that some words occur only in conjunction with others, and that words within a set of synonyms may co-occur with different collocates. The occurrence of one word alongside another is largely determined by meaning, but is not always predictable from the meanings of the associated words alone. Dholuo provides a clear instance in the lexical field of brownness, where three distinct lexemes are each restricted to a particular class of referent.

(1) Nyako malando dhi dala.

‘The brown [light-skinned] girl is going home.’

(2) Dhoot ma rabuor otur.

‘The brown door is broken.’

(3) Wuowi ma silwal riembo ndiga.

‘The brown boy is riding a bicycle.’

The three lexemes malando, rabuor and silwal all gloss as brown in English, yet they are not interchangeable: each selects a different class of collocate. This is the strongest evidence in the data for Palmer’s claim that collocational restriction, rather than denotation, differentiates members of a synonym set, and it is also the point at which an English gloss is least informative, since English brown neutralises a distinction that Dholuo maintains.

The restriction appears to be governed by the class of referent rather than by any difference in hue. Malando is selected for human complexion, rabuor for inanimate objects, and silwal for a third class of referent distinguished in the data. The substitution of one for another outside its proper class is rejected by speakers, which confirms that the differentiation is collocational rather than denotational and that it is invisible in the English gloss.

4.2Homonymy

A lexeme constitutes a case of homonymy if it shares distinctive properties with another, namely grammatical category and grammatical properties, sound form and spelling, while carrying unrelated meanings (Löbner, 2002). In Dholuo, homonymy is realised across different grammatical categories, and the context in which the word occurs determines the implied meaning, since items with unrelated meanings would otherwise create ambiguity. The form yien illustrates the pattern.

(4) Isekelo yien mar chweko?

‘Have you brought firewood?’

(5) Isekelo yien mar amuonya?

‘Have you brought tablets?’

(6) Isekelo yien ma ipidho?

‘Have you brought tree seedlings?’

(7) Isekelo yien mar ajuoga?

‘Have you brought the charm?’

In each case the interpretation of yien is fixed by the modifying phrase that follows it, so that disambiguation is achieved syntagmatically rather than by the form itself.

The classification of yien merits comment, since it bears on how strictly the relatedness criterion is to be construed. The four senses can be traced to a common association with trees and plant material, and on a broad reading of relatedness the item would constitute polysemy by metonymic extension rather than homonymy. The present analysis treats it as homonymy on the narrower criterion adopted from Löbner (2002), namely that the senses are not synchronically related for speakers, who report no felt connection between the firewood sense and the charm sense. The case illustrates that the boundary between homonymy and polysemy in Dholuo is determined by speakers’ synchronic intuitions rather than by etymology alone.

4.3Polysemy

Polysemy may be defined as one form, written or spoken, having multiple meanings that are all related by extension (Yule, 1996). Difficulties sometimes arise in distinguishing polysemy from homonymy. Polysemous lexemes are etymologically related and typically originate in metaphorical usage.

4.3.1Metaphorical extension

The lexeme ich, whose central sense is ‘stomach’, acquires a further sense related to the first but used figuratively.

(8) Ich kao dweye ochiko.

‘A pregnancy takes nine months.’

A comparable extension operates on jakwath, whose central sense is ‘shepherd’.

(9) Jakwath biro lemonwa.

‘The pastor will pray for us.’

In both cases the extended sense is derived by metaphor from the central sense: the container for its contents in the case of ich, and the tending of livestock for the pastoral care of a congregation in the case of jakwath. The latter also illustrates how religious vocabulary enters the lexicon through semantic extension of existing items rather than through borrowing.

4.3.2Shifts in application

Ullmann (1970) notes that words have a number of different aspects of meaning according to the contexts in which they are used; some of these aspects are purely ephemeral, while others develop into permanent shades of meaning, and as the gap between them widens they may eventually be regarded as distinct senses of the same term. He further cautions that word classes other than the noun are equally exposed to such shifts. The verb winjo is polysemous in precisely this way, its senses distinguished by the class of complement it takes.

(10) Awinjo koyo.

‘I am feeling cold.’

Awinjo koko.

‘I have heard a noise.’

Awinjo thum.

‘I am listening to music.’

Awinjo gima iwacho.

‘I have understood what you said.’

The four senses of winjo — physical sensation, auditory perception, attentive listening and comprehension — form a graded series rather than a set of discrete meanings, and the sense selected on any occasion is determined by the complement. This is a case of broadening in the lexical pragmatic sense, the encoded concept of perception being adjusted in context to a more specific concept appropriate to the complement.

4.3.3Register

Polysemy in Dholuo also arises from the use of a word in a particular trade or profession. Bréal, quoted in Ullmann (1970), observes that in every trade or profession there is a certain idea so constantly present to the mind, and so clearly implied, that it seems unnecessary to state it when speaking. The lexeme waro, whose general sense is ‘relieve’, has specialised senses in restricted fields.

(11) Omollo nowara higa mokalo.

‘Omollo bailed me out last year. (legal register: ‘bail’)’

Yesu nowara.

‘Jesus saved me. (religious register: ‘save, redeem’)’

The two readings share the general concept of relieving a person from a difficulty, narrowed in each case by the register in which the utterance occurs. Register-conditioned narrowing of this kind is not recoverable from the lexical entry alone and provides further support for a pragmatic account of sense.

5Conclusion and recommendations

This study set out to identify the sense relations obtaining in Dholuo within a lexical pragmatic framework. Three relations were established: synonymy, homonymy and polysemy. The analysis showed that these relations are realised along several dimensions, namely formality, dialectal difference, borrowing, collocation and speakers’ evaluative meaning, and that the interpretation of items standing in these relations depends on pragmatic processes of narrowing and broadening rather than on encoded meaning alone.

Three findings merit emphasis. First, borrowed lexemes in Dholuo come closest to absolute synonymy, since borrowed and inherited items may share denotation, register and collocational range. Second, collocational restriction emerged as the principal factor differentiating synonyms that are otherwise denotationally equivalent, the field of brownness providing the clearest instance. Third, the data support Mugure’s (2009) conclusion for Gikuyu that lexical pragmatic theory handles polysemy and homonymy more readily than synonymy, since the dimensions along which Dholuo synonyms differ are heterogeneous and not all reducible to narrowing or broadening.

The findings are of use to students and professionals in linguistics, particularly those working on Nilotic languages, and to lexicographers, since the dimensions of synonymy identified here bear directly on how dictionary entries for Dholuo should distinguish near-synonyms. Further research should establish whether comparable relations and semantic features obtain in other Western Nilotic languages such as Acholi, Alur and Lango, and should extend the present analysis to antonymy, hyponymy and meronymy, which fall outside the scope of this study.

Declarations

Conflict of interest. The authors declare no conflict of interest.

Funding. This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

1. [1]Atoh FO (2002). Semantic analysis of Dholuo nouns: The semantic field approach [Unpublished M.A. dissertation]. University of Nairobi.
2. [2]Blutner R (1998). Lexical pragmatics. Journal of Semantics, 15(2), 115–162.
3. [3]Carston R (2002). Thoughts and utterances: The pragmatics of explicit communication. Blackwell.
4. [4]Cruse DA (1986). Lexical semantics. Cambridge University Press.
5. [5]Cruse DA (2000). Meaning in language: An introduction to semantics and pragmatics. Oxford University Press.
6. [6]Crystal D (1997). The Cambridge encyclopedia of language (2nd ed.). Cambridge University Press.
7. [7]Löbner S (2002). Understanding semantics. Arnold.
8. [8]Lyons J (1968). Introduction to theoretical linguistics. Cambridge University Press.
9. [9]Lyons J (1981). Language and linguistics: An introduction. Cambridge University Press.
10. [10]Lyons J (1995). Linguistic semantics: An introduction. Cambridge University Press.
11. [11]Mugure CM (2009). A lexical pragmatic analysis of sense relations in Gikuyu [Unpublished M.A. thesis]. University of Nairobi.
12. [12]Murphy ML (2003). Semantic relations and the lexicon: Antonymy, synonymy and other paradigms. Cambridge University Press.
13. [13]Mwebia FG (2006). A lexical pragmatic analysis of sense relations in Kimeru [Unpublished M.A. thesis]. University of Nairobi.
14. [14]Ndlovu E (2001). Sense relations in the treatment of meaning in Isichazamazwi SesiNdebele. University of Zimbabwe.
15. [15]Oduol JA (1990). Dholuo dialects: Synchronic state and historical inference [Unpublished Ph.D. dissertation]. University of Nairobi.
16. [16]Oduor JAN (2002). A study of syllable weight and its effects in Dholuo phonology [Unpublished Ph.D. thesis]. University of Nairobi.
17. [17]Okombo DO (1982). Dholuo morphophonemics in a generative phonology framework. Dietrich Reimer Verlag.
18. [18]Omondi LN (1982). The major syntactic structures of Dholuo. Dietrich Reimer Verlag.
19. [19]Palmer FR (1976). Semantics: A new outline. Cambridge University Press.
20. [20]Palmer FR (1981). Semantics (2nd ed.). Cambridge University Press.
21. [21]Saeed JI (2003). Semantics (2nd ed.). Blackwell.
22. [22]Stafford RL (1967). An elementary Luo grammar, with vocabularies. Oxford University Press.
23. [23]Sweetser E (1991). From etymology to pragmatics: Metaphorical and cultural aspects of semantic structure. Cambridge University Press.
24. [24]Ullmann S (1970). Semantics: An introduction to the science of meaning. Blackwell.
25. [25]Wilson D (2003). Relevance and lexical pragmatics. Italian Journal of Linguistics, 15(2), 273–291.
26. [26]Yule G (1996). The study of language (2nd ed.). Cambridge University Press.