This paper investigates the division of labour 1 operative in naturally occurring bilingual discourse. It attempts to establish the grammar of intersentential codeswitching (ISC) in mixed Yoruba-English interrogative sentences. The paper seeks to find out what constitutes the asymmetry between the participating languages (Yoruba and English), as they supply important grammatical structures in order to produce a well-formed mixed YorubaEnglish interrogative sentences. Asymmetry in the division of labour is observed in two ways: first between the type of morpheme involved in the switch (content or system morpheme, depending on whether they assign/ receive thematic roles or they do not), and second, how the morphemes are assigned roles and this depends on which language is the Matrix Language (ML) and which is the Embedded Language (EL). The paper observes that in YorubaEnglish bilingual speech, when uniformity of structure is juxtaposed with asymmetry, structures of the ML are preferred according to the MLF principles (Myers-Scotton 1993; and Myers-Scotton 2011). The paper argues that even in this not-so-much investigated area of codeswitching in mixed Yoruba-English sentences, in the code switching structure, the ML is invariably Yoruba and as a consequence determines the permissible and non-permissible combinations for well-formed structures, contrary to what some scholars have claimed.
Languages in contact consequently influence each other. Such influence is felt in a number of ways; one is borrowing and another is codeswitching (CS). Borrowing is usually restricted to vocabulary or lexical items. Even so, the distinction between borrowing and codeswitching, no matter how thin, should be established. Field (2002: 3) says,
Owino (2003: 26), quoting Crystal (1987) explains that borrowing involves introducing a word or other linguistic feature from one language to the other. Loanwords, he says, are vocabulary borrowings which fill
According to Poplack (1980), CS follows a set of grammatical rules of the languages of interaction, and any departure from them makes the sentence unacceptable. CS is a process where the initiator of speech or the speaker switches from one code or language to another as conditioned by the situation, audience or subject matter ( Essien. O .1995).
Field (2002: 184,185) describes codeswitching this way:
CS and codemixing(CM) have at times been distinguished. Essien (1995:272) defines CM
A closer look at Poplack’s typology shows that only two types of CS really exists, inter-sentential CS and intra-sentential (CS). For example, Poplack’s tag-switching is a type of intra-sentential CS since it involves lexical insertion. More explanation is provided later in this paper on inter-sentential and intra-sentential CS types.
Bokamba (1989), cited in Amuda (2006:91) distinguishes between codeswitching and codemixing thus:
To see the point of Bokamba, consider the examples below. They illustrate how CM has been explained. The examples contain language mixing between Ibibio and English, Igbo and English and Yoruba and English 2 .
In the examples above, (1) is Ibibio-English codemixing in which the words
Myers-Scotton sees no distinctions between CS and CM. She however distinguishes between types of CS: intra-sentential CS and inter-sentential CS. We provide example of intra-sentential switching between Spanish and English in (4) below, taken from Myers-Scotton (2006:240):
This example is not different from (1) to (3) above. Myers-Scotton (2006) explains that (4) above is an example of a specific type of
In the (5), a case of inter-sentential or inter-clause codeswitching is found in the father’s speech,
Interestingly, Sridhar (1996) makes it clear that whereas codemixing occurs intra-sententially; code-switching occurs inter-sententially. Consequently, in this paper, the single term, code switching, is used for both cases. Following Myers-Scotton (2006), this paper uses the term CS to cover both so-called CM (lexical inserts) and CS (phrasal inserts), using Yoruba-English interrogative sentences, within the framework of using the Matrix Language Frame (MLF). CS is a case of language alternation. Myers-Scotton’s view of CS adopted in this paper can be diagrammed:
This is the CS view adopted in this paper. To us, CS is a consequence of language contact; it is language alternation. CS subdivides into intersentential CS and intra-sentential CS. The latter subcategorises into phrasal inserts and single-lexeme inserts. Thus all cases of so-called CM are actually example of CS. The rest of the paper is arranged as follows: immediately after this section, we look at the Matrix Language Frame (MLF) and the principles underlining it. Thereafter, we review CS literature in Yoruba-English mixed sentences. Following this, we present the data for the study for analysis. We then conclude.
1This social science term was borrowed from Myers’-Scotton (2009: 81) To us, this metaphor beautifully describes what operates in codeswitching when two languages are used in a discourse to supply the morpho-syntactic structure. 2Yoruba belongs to the Kwa family of the Niger-Congo phylum in West Africa. It is one of the three major indigenous languages spoken in Nigeria; the other two are Igbo and Hausa. Ibibio is a minority language in Nigeria.
The Matrix Language Frame claims that there is a
According to the MOP, the ML determines the order of the elements in ML + EL constituents. The SMP requires that content morphemes can only be drawn from the ML, given the distinction that the model makes between
Since its first proposal, the MLF has undergone modifications, and two of the most recent are in Myer-Scotton (2009, 2011) where she makes clearer the unique roles of the MOP and SMP as being the testable hypothesis at the heart of MLF model. She provides the roles below:
It should be made clear that the System Morpheme Principle does not in any way imply that
The 4- Model of morpheme classification refines the System Morpheme and adds precision to the MLF model. It identifies four ways by which morphemes are identified:
In summary, the MLF is based on three premises:
Despite its popularity as a model of CS analysis, MLF model has been attacked. Scholars like MacSwan (2000,2005a&b), van Gelderen & MacSwan, 2008) have argued that the CS is lexically-based and does not necessarily call for an ML. Specifically, MacSwan (2000: 43) contends that,
3The Complementizer Phrase (CP) is the highest unit projection in by lexical items, i.e. the highest level in a tree of syntactic structures (e.g. NP, VP, etc). It is headed by the Complementizer (COMP. e.g. that, if, whether, for). Myers-Scotton (2002: 54) defines the CP as “a syntactic structure expressing the predicate-argument structure of a clause, plus the additional syntactic structures needed to encode discourse-relevant structure and logical form of that clause”. It is argued that the CP is a more exact / unambiguous unit of analysis than both the “clause” and the “sentence”.
The earlier studies on the contact between English and Yoruba focus on both sociolinguistics and grammatical aspects in relation to the influence of English on Yoruba. Salami (1969) discussed loan/borrowing, and distinguishes between fully and partially assimilated English loan words. Akere (1980) investigated CS strategies of Ijebu, a dialect of Yoruba, using Ikòròdú 4 as his focal point. In his study, he found that even in community meetings where the indigenous language was supposed to be used, CS was resorted to. He argued that such verbal strategy was predicated upon by the
Scholars’ attentions have been drawn to the syntactic aspects of YorubaEnglish CS. Banjo (1983, 1996) and more recently, Lamidi (2003) have observed that Yorùbá–English CS, functional elements such as determiners, inflection elements such as tense, modal, aspect, and agreement, from English grammar are barred in favour of those from Yorùbá. These scholars argue that nouns can occur in either English or Yoruba, if there are multiple lexical heads; but the functional heads will be invariably Yoruba. Ayeomoni (2006), in his analysis of language use in Yoruba speech community observes that the grammatical class of code switched items of his study was always nouns. He further observes that the phonological shapes of such words
Lamidi (2004, 2008, and 2009) chose to do his CS analysis within Generative Syntax. In his analyses, Lamidi looks at switches at the word boundary, morpheme boundary, within the Determiner Phrase constituents, the serial verb constructions as well as at sentence and clause boundaries. Lamidi (2004, 2008 and 2009) observe that functional heads in Yoruba such as ‘pé’ and ‘kí’ belong to the class of complementizers. In Yoruba-English CS, they determine whether or not certain elements can be switched from one language to the other. He also observes that English personal pronouns (subject) are barred from occurring in Yoruba-English CS * we do not dwell on this in this paper. Even so, two of his examples on the use of ‘kí’ in Yoruba-English CS are of interest to us. Consider ‘a’ and ‘b’ below.
As in example 2 below, examples ‘a’ and ‘b’ are contracted in ‘c’ and ‘d’ to prove that Yoruba is the ML in the CS corpora. This is the usual case in informal speech, the domain of CS. More is said about ML below
Bamgbose (2004) focuses on the sociolinguistic construct of linguistic imperialism when he concludes that the contact between English and African languages such as Yoruba is regrettably characterized by increased power and prestige of English at the expense of other languages, a consequence of the spread and domination of Anglo-American culture and unfortunate positive attitudes and preference for English at the expense of one‘s own language. Bamgbose’s study does not present any serious linguistic analysis; it is purely sociolinguistic in scope.
This study sets out to add to investigate the grammatical patterning of CS in mixed Yoruba-English naturally occurring sentences through the lenses of the MLF model proposed by Myers-Scotton. We intend to present the extent to which the principles underlying MLF model is operative in the CS discourse of Yoruba-English bilinguals. To the best of our knowledge, scholars have not singled out the aspect of CS for investigation. Thus this is an unusual area of study and a good addition to the earlier study on YorubaEnglish CS.
We observe that the interrogator in mixed Yoruba-English intra-clause discourse, our focus in this study, is invariably Yoruba, and that the interrogator is usually blocked from occurring in English whenever the two languages are codeswitched. In other words, in such sentences, it seems that the Yoruba language determines the grammatical structure of the discourse.
To this end, the study intends to answer the following two research questions:
4One of the major towns in Lagos State, Nigeria, West Africa.
Interrogative sentences are sentences used to ask questions. They are used for requests. Depending on the questioner’s intention, a yes-or-no answer may be given or a full answer supplied. Auxiliaries, for example, have syntactic properties which make them amenable to inversion unlike the main verb (e.g it is a dog> Is it a dog?), with the auxiliary coming first. This seems to be the situation in most languages of the world. And going by the recommendation of Generative syntax, it is part of the UG (Chomsky (2013). Syntactic operations in Wh-word 5 is even more interesting where
Using the ML frame, this study investigates how auxiliary verbs and wh-words operate in mixed Yoruba-English interrogative sentences. We begin our discussion with the yes/no questions with auxiliary verbs. We observe that the auxiliary interrogator which triggers 6 the switch is invariably Yoruba’s. The examples are taken from naturally occurring sentences either recorded by the researchers of taken from the works of earlier researchers.
5A Wh-word is a question word like who/where/when/ etc beginning with wh (Radford, 2009) 6The idea of trigger (s)/triggering is borrowed from Clyne (1967) In triggering, according to Clyne, an item of ambiguous affiliation (i.e., one belonging to the speaker’s two languages) triggers off the switch from one language to the other.
When Yoruba is the ML in interrogative CS corpora, thus supplying the question maker, the interrogative morpheme (a system morpheme) changes 7 , depending on the presence or absence of the progressive marker ‘n’. If the progressive marker is present (see example 1), then a full NP, with the head from the EL and the specifier from the ML (by movement rule), is usually allowed immediately after the interrogator as its complement. After the interrogator, the ML supplies the VP; elements from the EL are disallowed at this point, except the second NP ‘bọ̀’ is deleted 8 and replaced with an element from either EL or ML. Some question markers cannot be the INF, of the IP in mixed Yoruba-English sentences (see footnote). Does this make English the ML 9 determining the morpheme structure? A surface reading may yield such an answer, but remember that ‘boy’ is only an insertion within the grammatical frame of another language. Besides, modern usage by new generation of native speakers of Yoruba has proved that Yoruba is the ML. For example, all the question markers, except ‘Ǹjẹ́’ (for stylistic reasons) can be moved to the end of the sentence but the inserted nominal element from the El cannot (consider ‘a’ and ‘b’). In addition, both content and system morphemes come from Yoruba, the base language. Of course, a content morpheme, ‘boy’ comes from English. But this does simply confirm the potency of the MLF model. Myers-Scotton (2009: 85) says,
Even when there is no progressive form as in the response of SB (student B), the fact of ML/EL dichotomy is not compromised.
Now, consider example 2 where ‘Șé’ an IP, changes from the copula to the dummy verb, DO. It seems that in mixed interrogative Yoruba-English sentences, ‘Șé’ assumes two distinct forms (copula and dummy), depending on the nature of its complement VP. It seems that ‘Șé’ does not become a copula except it has a possessive VP as the C of the IP (that is has/have/ need), and at times the sentence ends in a content morpheme from the EL (see example 2)
Setting: in a restaurant at Adeniran Ogunsanya College of Education (AOCOED), Lagos, Nigeria. It is evening, Females students discussing after a meal; perhaps ruminating over what had happened earlier during the day.
Compare the above with (a & b) below:
Setting: School campus (AOCOED), two students, a male and a female conversing on helping each other.
In examples 1 and 2 above, each of the constituents follows the Yoruba, not the English, word order. This example supports the Morpheme Order Principle (MOP) of the MLF model because these constituents and everything else in the clauses follow Yoruba order, indicating that only one language supplies morphemes order. This indentifies Yoruba as the ML. Again, in (example 2), the auxiliary HAVE (ní) showing possession comes from Yoruba and more significantly, this morpheme is attached to the noun ‘account’, an English lexical constituent, to form a structure which obeys the rules of Yoruba grammatical frame where the deletion 11 process in highly observed in order to avoid hiatus for well-formedness (normally, the HT 12 will have become a floating tone put on the next vowel if the word ‘account’ were Yoruba) to be realized. Of course, the possessive verb ‘ní’ may be separated from account (e.g. ní
Note also that the MT appendage to the adjective ‘fine’ in (v) is necessary for the insertion of the verb it precedes. The point is that a feeding rule process operative in Yoruba in which phonological processes take place in succession is observed. First, the verbal element is deleted; then its HT delinks paving way for the succeeding LT in order to insert the English verb, ‘understand’. Take note too of the second person plural pronoun, used for a singular referent. This is an honorific pronoun used in the plural form in Yoruba for singular referent for the purpose of honoring the referent. These instances support the system Morpheme Principle that requires such morphemes to come from only one of the participating language. Lastly, the question trigger is invariably the Yoruba close-ended yes-no polar interrogative trigger. This too marks Yoruba as the ML in all the clauses.
We explained above that the MLF model provides the morphosyntactic frame for the sentence, within which lexical items from the EL may occur. All this takes place in what is called the
The examples analysed above therefore prove that Yoruba is the ML and not otherwise, and that should the order be reversed, the sentences become ungrammatical and ill-formed. Compare the two examples below- the second is ill-formed when the Morpheme Order Principle is tampered with:
Kí ló ń
QM BE PRG.M happen friend QM +2S love credit account 1sg FOC
Even though the single example here may be syntactically ‘well-formed’, it is unusual and definitely substandard, even by L2 users of English!
7The question marker can be any of the following: Șé/Ǹjẹ́/Àbí/ Bẹ́ ẹ̀ as in the example: But when the progressive marker ń bearing a HT is present, some question markers are disallowed from surfacing. Consider: Note: The categorical status of the MT ‘ni’ in Yorùbá́ syntax has continued to be controversial. Bowen (1858) referred to it both as a verb and as a particle; other scholars including Abraham (1958), Delano (1958), in Awobuluyi (1992) have called it a verb, a marker, particle and expletive. Awobuluyi (1992) says it is a focus marker. More recently, Adesola (2006) sees it as a particle while Ilori (2010) says it is a focus marker. The controversy is far from being over! To us in this paper, ‘ni’ is focus marker, an emphasize. 8An example is: Șé/Ǹjẹ́/ man yẹn ń care fun/ tọ́jú) e? 9See Lamidi (2004, 2008, 2009 and 2013) for a thorough discussion on how Yoruba has consistently been the ML in a mixed Yoruba-English sentence. The only exception cited by Lamidi, 2013: 325):Anything to fenu gbé? Anything to use + mouth carry ‘Anything to eat?’ 1010 Șé aunt yẹn ń padà bọ̀?Becoming:Ș’aunt yẹn ń padà bọ̀? 11For a discussion of how deletion operates in Yorùbá́, especially in loan words, see Akinlabi (1993) and Orie (2012) 12Yoruba is a three-toned language: High Tone (HT in wá= come), Mid Tone (MT in wà) and a Low Tone (LT, usually unmarked). For a discussion on Yoruba tonal system, see Akinlabi (2004)
Wh-question is another area where the power of ML over EL is attested as shown in the following example:
When Yoruba is the ML in CS corpora, there is no trouble retaining double determiners and double prepositions (see Example 3). In example 3, two prepositions (grammatically required null preposition, after ‘
The case with the preposition is similar. The second preposition is easily found in English structures; it is the head of the PP, ‘nípa Prof’, ‘
But what about the first preposition? The first preposition is a special preposition, required for well-formedness, which may at times surface (as in ‘ní ilé…’ with other parts of the sentence in example 3 intact) or remain unseen as a lexical item as it is in example3. It is the preposition ‘ni’. ‘It will be good to note that the preposition, ‘ní’ in Yoruba is locative; it means, ‘in’ or ‘at’. About it, Folarin Schleicher (2008: 12) says: On the other hand, the preposition
But there is more to this preposition. Significantly, Schleicher’s book is meant for the inexperienced learners of the Yoruba language, hence is silent about the nuances of usage of this preposition. What is its purpose in the CS corpora? It seems it is there to add some grammatical information to the noun. It seems to be emphasizing that the speaker never expected what happened, going by a previous discussion (see the response of speaker B and then the conclusion of speaker A). This is further confirmed by the focus marker 14 ‘ni’ removing any doubt as to where the man was found, ‘in’ not ‘beside’ her house.
Recgnizing that the two determiners, and especially the two prepositions have different functions is necessary to defeat any analysis of redundancy. This is also critical to the application of MLF. This also proves that Yoruba is the ML, not English.
13Myers-Scotton (2009: 87) observes a similar thing in CS with Arabic as the ML and English, Dutch, French, etc. as EL. Whereas Myer-Scotton’s observes that‘a pronoun from Arabic seems to double a pronoun from the EL’, the doubling in this paper does not affect the EL in the same way. It seems to be the normal practice in the language. 14Note the difference between ‘ní’ with a HT and ‘ni’ with a MT. The later has always been controversial as indicated above. Following Ilori (2010) and Awobuluyi (1992), we call the latter a focus maker.
Contrary to the argument of MacSwan and his associates that there is no ML in CS, there is an active operation of the MLF in and the application of its principles to Yoruba-English mixed constituents. We proved that the CS structure is highly governed by the Yoruba grammatical frame so much so that any attempt to soften the construction will produce ungrammaticality. Like others before it, this study also confirms the universal application of Myer-Scotton’s MLF model and Classic Code switching.
Intuitively, as native speakers of Yoruba who use the language for everyday linguistic activities, and from our observation of the speech of fellow native speakers of the language, as indicated by the data presented in this study, we can conclude that the MLF model has really become by far the one of the most popularity theories of CS. This study has proved that as far as Yoruba-English mixed constituents are concerned, there is an ML-EL dichotomy in bilingual speech.
It is likely that new pattern of CS with Yoruba as one of the participating languages will emerge. Evidence indicates that they are already emerging as we see with the example in Lamidi (2013:325, 326). He himself said,
QM: Question Marker
BE: any form of the BE verb
PL.DM: Plural Demonstrative
1,2,3: 1st person,2nd person, 3rd person
MOD: Modal Auxiliary
PRG.M: Progressive Marker
INT.M: Interrogative Marker
PRG.M: progressive Marker
NOM.: Nominative Case
ACC: Accusative Case
SA: Student A
SB: Student B