South Africa has 12 official languages: isiZulu, isiXhosa, siSwati, isiNdebele, Sesotho, Setswana, Tshivenda, Xitsonga, Sepedi, English, Afrikaans, and South African Sign Language. Although isiZulu and Xitsonga belong to different subgroups, they are nonetheless part of the Bantu languages family prevalent in South Africa. According to Herbert and Bailey (2002:50), isiZulu is classified as a Bantu language within the Nguni subgroup and belongs to the Niger-Congo language family. With 14 613 202 native speakers, accounting for 24.4% of the population, isiZulu is recognised as the most widely spoken language in the country, predominantly in KwaZulu-Natal (Statistics South Africa 2022). In contrast, Xitsonga, often referred to as ‘Tsonga’, is spoken by 2 784 279 individuals, representing 4.7% of South Africa’s population (Stats SA 2022). Most Xitsonga speakers reside in the provinces of Limpopo, Mpumalanga and Gauteng, with Limpopo being the primary area of usage. IsiZulu and Xitsonga exhibit certain linguistic similarities, such as their subject-verb-object (SVO) structure and a complex agreement system (Keet & Khumalo 2017:131; Lee, Babane & Hlungwani 2022:2–4). Nevertheless, they also possess unique characteristics shaped by their distinct cultural contexts.

Substantial gaps remain in comparative studies of these languages’ morphosyntactic structures, specifically in relation to universal quantifiers. Existing studies on these two languages looked at grammatical categories, such as verbalisation patterns for existential and universal quantification (Keet & Khumalo 2017) or syntactic structures such as pre-nominal DP modifiers and penultimate lengthening (Lee & Riedel 2021). While Zerbian and Krifka (2008) studied quantification in Bantu, they did not focus on Xitsonga. Therefore, comparative studies on the morphosyntactic behaviour of universal quantifiers in the two Bantu languages have received limited attention, despite these languages’ complex and flexible agreement systems. It is posited that the intricate agreement systems of class nouns in both languages contribute to the development of universal quantifiers (Masonto 2024:150). This is because universal quantifiers agree with their head nouns in noun classes.

This article aims to analyse and compare the morphosyntactic features of isiZulu and Xitsonga and examine the functional roles and distribution of universal quantifiers within these languages. It sought to answer the following research questions: (1) How are universal quantifiers formed in these languages? (2) How do they function within phrases or sentences? (3) What are the morphological and syntactical similarities and differences between these universal quantifiers?

Noun phrases (NPs) in Bantu languages display distinctive structural and syntactic characteristics. Typically, an NP consists of a head noun accompanied by various dependents, including possessives, demonstratives, adjectives, numerals and quantifiers. The arrangement of these elements can vary significantly among different Bantu languages. For instance, while many languages follow a structure where the head noun precedes its dependents, others may exhibit different patterns (Achiri-Taboh 2023:1; Lusekelo 2009:316; Rugemalira 2007:137). Radford (2004:34–35) characterises quantifiers as functional categories that do not provide explicit descriptive information and can modify any semantic noun class that is not constrained by grammatical rules. As a subset of determiners, quantifiers enhance noun phrases by conveying quantificational attributes. Zerbian and Krifka (2008) identify three groups of quantifiers, namely intersective, universal and the class of proportionality. Specifically, universal quantifiers modify nouns by co-occurring with them and other elements to signify totality (Cushman 2019:240; Zerbian & Krifka 2008:12). The application of universal quantifiers are examined by presenting various examples (see Examples 1-14). All abbreviations used in the universal quantifiers examples are explained in Box 1.

In example 1, the universal quantifier ‘wote’ co-occurs and succeeds the noun ‘watoto’ to quantify it. It also agrees with ‘watoto’ by copying its prefix (wa-). In some Bantu languages, such as isiZulu, universal quantifiers typically precede nouns during quantification (Keet & Khumalo 2016:178). Quantification refers to the expression of quantity with language structure (Iordăchioaia 2019) as is presented in Example 2 (Keet & Khumalo 2014:5).

Conversely, Swahili also adopts a distinct system for universal quantifiers. The Swahili quantifier ‘kila’ [every] consistently precedes the noun it modifies, yet it is notable for its disagreement with the noun’s class features presented in Example 3 (Zerbian & Krifka 2008:12–14).

In this instance, ‘kila’ is positioned before ‘mottos’ without altering its form to correspond with the noun class. In contrast, the Swahili universal quantifier ‘-ote’ [all] functions differently, following the noun and agreeing with it, as illustrated in Example 4.

Here, ‘-ote’ adopts the prefix ‘wa-’ from ‘watoto’, modifying its form to align with the noun. Furthermore, Northern Sotho exhibits a pattern that is both similar and unique. The universal quantifier root ‘-ohle’ [all or whole] aligns with its head noun, reflecting its prefix as shown in Example 5 (Mojapelo 2007:110).

‘-ohle’ corresponds to the prefix ‘lo-’ from ‘lenao’. However, not all universal quantifiers in Northern Sotho adhere to this structure. The expression ‘ka moka’ [all or whole] does not exhibit agreement with its head noun, akin to the usage of ‘kila’ in Swahili, illustrated with Example 6.

Furthermore, Zerbian and Krifka (2008:15) also contend that in Northern Sotho, ‘every’ is communicated by the collective phrase ‘-ngwe le -ngwe’ [a certain, some, other], in which the quantifier ‘-ngwe’ agrees with the quantified noun using both nominal and pronominal agreement. This is demonstrated with Example 7.

Moreover, the Loogori language quantifier ‘vuri’ [every] is essentially distributive, while ‘-oosi’ is not (that is it permits distributive or collective interpretations), like in Example 8 (Landman 2016:220).

‘vuri’ is distributive in nature, yet it is not collective. In instances where a collective reading is forced, the phrase is ungrammatical. This is illustrated in Example 9 (Landman 2016:221).

Landman (2016:226) suggests that it is typical for ‘vuri’ to combine with singular count nouns if they express characteristics of single entities. Semantic adjustment would lead to specific readings when the quantifier ‘vuri’ pairs with plurals or mass nouns (which, by general assumptions, do not represent single entities). It is anticipated that ‘vuri’ must appear before the noun, similar to Swahili ‘-ote’ in the example above, given that the determiner phrase (DP) is head-initial, as phrase structures in Logoori generally are. ‘Vuri’ may not co-occur with pronouns since they share the same position.

In the Kihehe language of Tanzania, the universal quantifier ‘mbe-AGR2-li’ presents another fascinating variation. This quantifier serves both collective and distributive functions, consistently aligning with the noun according to its class (Cushman 2019:240). Whether it modifies a plural noun, as in ‘vá-nu mbẹ́-va-li’ [all people], or a singular noun, as in ‘m̩-víli mbé-gu-li’ [a or the whole body], ‘mbe-AGR2-li’ maintains agreement with the noun. The components ‘mbe-’ and ‘ -li’ lack independent meaning; they only convey significance within the complete quantifier, like the role of ‘moka’ in Northern Sotho’s ‘ka moka’. This illustrates the essential nature of the pre-prefix ‘mbe-’ in ‘mbe-AGR2-li’, which is grammatically incomplete without it (Cushman 2019:241).

In the cases of isiZulu and Xitsonga, universal quantifiers frequently display comparable patterns of agreement and positioning in relation to the nouns they modify. In isiZulu, the notion of ‘all’ or ‘each’ is conveyed through the stem ‘-onke’, which is accompanied by a noun class marker as Example 10 illustrates (Keet & Khumalo 2017:11).

Here, the universal quantifiers are positioned before their respective nouns, yet they do not replicate the nouns themselves. Conversely, in Xitsonga, the universal quantifier stem ‘hinkw-’ (all/the whole of/the entire) also succeeds the noun but does replicate its prefix, as illustrated in Example 11 (Lee & Riedel 2021:109).

These constructions reveal a consistent pattern of agreement between universal quantifiers and their corresponding nouns. The differences observed among Bantu languages, particularly isiZulu and Xitsonga, emphasise the intricate nature of quantification, with each language adhering to its distinct set of rules and exceptions. This variety highlights the significance of examining these specific languages (isiZulu & Xitsonga).

This study is grounded in Chomsky’s (1957) Transformational Generative Grammar (TGG) theory, which explains the syntactic structures underlying natural language. The theory postulates that sentences are developed using a set of formal rules producing two syntactic levels of representation: the deep structure and the surface structure. The deep structure describes the underlying grammatical relationships, while the surface structure shows the spoken or written form (Perlmutter 1968:9). Transformational Generative Grammar describes the way in which transformations function in deep structures in order to produce acceptable surface sentences. This study adopts the concepts of ‘deep structure’ and ‘surface structure’ to describe how universal quantifiers are positioned and interpreted within phrases, clauses and sentences, particularly through transformational rules. In this case, syntactic elements such as quantifiers, especially universal quantifiers like all and every are generated in specific structural positions at the deep structure level and may undergo syntactic movement as the sentence transforms into its surface structure. This is illustrated in English with Example 12.

In Example 12, the universal quantifier ‘all’ is part of the noun phrase ‘all students’, which functions as the subject of the sentence. At the deep structure level, this universal quantifier determines the scope of the noun it modifies. If this sentence were rephrased into a passive or cleft construction, such as ‘It was all students who went home’, the quantifier remains semantically linked to the noun ‘students’ despite the rearrangement of sentence elements. This illustrates how transformational rules affect surface structure, while preserving the syntactic and semantic relationships established in deep structure. The concept is also referred to as ‘quantifier floating’ (Sportiche 1988). Quantifier floating is a syntactic phenomenon in which a quantifier is displaced from the noun phrase it modifies, while maintaining its semantic association (Al Khalaf 2019).

In the context of this study, the position and agreement of universal quantifier ‘-onke’ can be analysed through these principles. This is shown in Example 13.

At the deep structure level, sentences (Example 13a to Example 13b) originate from the same underlying proposition (all the people are going). The universal quantifier ‘bonke’ quantifies the noun ‘abantu’ [people], forming the quantified noun phrase ‘bonke abantu’ [all people]. However, the sentences differ in the surface structure, which reflects how the sentences are expressed after transformational rules (movement, reordering) have been applied. For instance, in Example 13a, the universal quantifier appears post-nominally. This is the canonical surface form and reflects the deep structure more directly; the quantified noun phrase remains intact. In Example 13b, ‘bonke’ appears post-nominally. This reordering is a surface phenomenon and does not alter the meaning. Example 13c demonstrates quantifier floating, ‘bonke’ appears dislocated from the noun it quantifies and is positioned after the verb phrase (bayahamba). This suggests that the quantified noun phrase (bonke abantu) has undergone movement, and the quantifier has been ‘floated’, while still semantically modifying ‘abantu’. This is also seen in the Xitsonga examples in Examples 14a and 14b.

Similarly, sentences (Examples 14a and 14b) above all mean ‘all the people are going’, showing variations in the surface structure. In Example 14a, ‘hinkwavo’ succeeds its noun, while it precedes it in Example 14b. In Example 14c, the quantifier is floated, but still quantifies the noun ‘vanhu’ [people]. In both languages, despite surface variation, the deep structure remains the same; the universal quantifier relates to the subject noun. These quantifiers are syntactically separated from their nouns, but they semantically modify them.

This study followed an interpretive paradigm to collect qualitative data from Bibles and language practitioners. In line with the chosen research design, a comparative research design (Mello 2021) has been chosen to allow the researchers to analyse and contrast the structure and distribution of universal quantifiers across the two languages. This article drew on one isiZulu and one Xitsonga Bible translation, specifically the 2020 isiZulu version Ibhayibheli Elingcwele and the 2012 Xitsonga version Mahungu Lamanene translated from the English Good News Bible. It also used purposive sampling to intentionally select data from four language practitioners (two from isiZulu, two from Xitsonga) and two Bibles. Bible translations were chosen for their standardised language and accessibility, making it easier to identify universal quantifier forms and consistent examples of universal quantifiers in isiZulu and Xitsonga. These texts served as a starting point for isolating relevant quantifier constructions. Interviews were then conducted to explore the morphosyntactic structure of these quantifiers, focusing on how noun class systems, concordial agreement, and syntactic position shape their use in natural speech. Additionally, engaging with language practitioners allowed for the collection of further examples and expert insights, enriching the data with real-world usage and grammatical explanation.

Researchers deliberately searched for occurrences of the stems ‘-onke’ and ‘hinkw-’ in the Bibles. Instead of selecting texts randomly, they focused only on Bible entries that contained words derived from these stems. Once the occurrences were identified, researchers extracted words, phrases, clauses and sentences that reflected universal quantifiers in isiZulu and Xitsonga. This method ensured a focused and in-depth analysis of universal quantifier positioning. Furthermore, four language practitioners were interviewed; these were purposely selected due to their language expertise and years of experience. Their profiles are outlined in Table 1.

During interviews, participants were presented with samples from the Bible, and asked to interpret the structures, meaning and requested to provide additional examples. The data from the Bible and interviews were then analysed through thematic analysis, and three themes emerged: the morphological variations of the universal quantifiers ‘-onke’ and ‘hinkw-’; the morphological and syntactical structures of ‘-onke’ and ‘hinkw-’ and the co-occurrence restriction of ‘-onke’ and ‘hinkw-’.

This section presents the findings of the study on the morphological variations of the universal quantifiers ‘-onke’ in isiZulu and ‘hinkw-’ in Xitsonga. Three themes are discussed.

This comparative study exposes the complex morphosyntactic patterns governing quantification in the two selected Bantu languages, specifically through the stems ‘-onke’ and ‘hinkw-’. While isiZulu uses prefixal agreement and restricts quantifier floating in singular forms, Xitsonga utilises suffixal agreement and permits more flexible positioning of quantifiers in both singular and plural constructions. The findings highlight the critical role of noun class agreement and demonstrate how universal quantifiers reflect both syntactic structure and semantic scope. This study significantly contributes to the field of African linguistics by addressing a gap in comparative morphosyntactic research, deepening our understanding of how agreement systems and quantifier behaviour manifest across related languages. It is recommended that future research expand to include other Bantu languages, explore dialectal variations, such as those found in Mozambican Xitsonga, and investigate the semantic distinctions between distributive and collective quantification.