More on the Categorial Status of ( T ) àbí in Yorùbá Grammar

Yorùbá language is one of the major languages spoken in Nigeria. The term is also used to refer to the language and the native speakers. As shown in Oyetade, Yorùbá language is spoken in six states that constitute the southwest of Nigeria – Lagos, O ̣ ̀ yo ̣ ́ , O ̣ ̀ ṣun, Ògùn, Òndó, and Èkìtì. This study investigated the Standard Yorùbá used in the Southwest Nigeria by focusing on the conjunction t(àbí). Findings reveal that there are varieties of Yorùbá language based on the location of the speakers and the state they occupy in Nigeria: Ọ̀yọ́ dialect, Ègbá dialect, Èkìtí dialect, Òndó dialect and Ọ̀wo ̣ ̀ dialect to mention a few. Previous scholarly works on Yorùbá grammar show that (t)àbí performs two functions and it is ascribed with two nomenclatures namely conjunction and polar question word. However, this present paper provides another view that is different from the views of the earlier scholars. Findings in this study reveal that t(àbí) is a conjunction in all its positions of occurrence and the researcher argues against its use as a polar question word. It is established in this study among other things that its occurrence at sentence initial position is as a result of ellipsis. The study also maintains that where it appears at sentence medial position, the polar question word has been deleted.

Conjunction and question words are integral parts of the universal grammar and every language has the two in its repertoire (lexicon).No language has been reported where the two are absolutely missing.Conjunction is a grammatical category that joins words, phrases, clauses and sentences together while polar question words in Yorùbá are interrogative particles that are adjoined to a declarative sentence to make it interrogative.The grammatical category (conjunction) and the interrogative particle (polar question) are used to show different syntactic structures.
However, over the years, Yorùbá grammarians like Bamgbose (1967Bamgbose ( , 1990)), Awobuluyi (1978), Ilori (2010) and Ajiboye (2013) have continued to ascribe two functions to (t)àbí from the traditional linguistic approach.None of these scholars, to the best of our knowledge, has queried the assertion that (t)àbí is a conjunction and a polar question word.It shall be argued in this paper that all occurrences of (t)àbí perform the function of a conjunction.It will also be shown that there is no single word that can combine the functions of a polar question and a conjunction.For instance, a polar question word serves as the introducer of interrogative sentence while a conjunction occurs between nouns, clauses and sentences.This significant difference is an indication that (t)àbí cannot perform the two functions simultaneously.
This paper is divided into four sections.Section one will form the introduction.In section two, the researcher discusses the previous analysis of (t)àbí as a conjunction and a polar question word.In section three, arguments will be presented to support the present proposal.Section four is the conclusion.
In the examples in 4 above, it is shown that nígbàwo "when" is the question word in 4 (a) and (t)abí is the conjunction in the sentence; while the same (t)àbí is the question word in 4 (b&c).A closer look at example 4 (a) will show that (t)àbí is a conjunction in both 4 (b&c) because a question word can be added to 4 (b) while the clause before (t)àbí in 4 (c) has been deleted.Furthermore, the question word in 4 (c) has been deleted.The researcher agrees with Ilori's claim in 5 (b & c) that (t)àbí is a clausal conjunction in interrogative construction but Ilori (2010) has not provided the question words in the sentences.He (Ilori) has not been able to explain the function of (t)àbí when it occupies a sentence initial position.This present study shows that examples 5 (b and c) can be rendered in these forms and it will be grammatical.If the sentences in 6 (a and b) are grammatical in Yoruba, the researcher suspected that the question words in 5 (b and c) have been deleted.Moreover, it is worth noting that (t)àbí joins sentences of equal status where one is assertive the other is negative.It is also observed that when the first clause is introduced by polar question words such as ṣ é and njẹ , the question word is always deleted in the second clause in Yorùbá utterances (see example 1 b and 2 b above), this, the researcher assumes must have made some scholars think that (t)àbí is the question word in the second sentence.However, if examples 6 (a and b) are grammatical; thus, the polar question in the second clause has undergone ellipses.Ajiboye (2013, p. 10) makes two assertions with respect to (t)àbí.First, he claims that there are three lexical items having the same segmental and tonal features in the form of (t)àbí but function differently; one function as a conjunction, the second as Yes-no interrogative particle and the third as a kind of Whword.Secondly, he notes further that the kind of interrogative particle that comes up depends on the syntactic context and the type of structure.Lastly, he assumes that (t)àbí is one and the same lexical item that demonstrates ambiguity between those possible meanings, and one can only appeal to context in order to determine which meaning is intended.Is it a fact that Ade went to the farm or he slept?
In reaction to Ajiboye's assertion that (t)àbí is a polar question in one context, a content question in another and a conjunction in some other environment, a closer look at the examples in 8 (a, b & c) above, will show that (t)àbí has a peculiar function which is a clausal conjunction.In addition, a content question in Yorubá has a strong [Wh-feature] that forbids any other word to replace it.Moreover, when polar question words are adjoined to where (t)àbí is purported to be a polar question word like the examples in 8 above what happens to (t)àbí?As the researcher has noted earlier, (t)àbí combines two sentences with equal structure (S1 conj S2), what happens in most cases is the deletion of the question word (Qw) and NP of S2 to avoid repetition.Observe that 8 c is deviant when the question word (Qw) is attached to S2; we suspend the explanation for this to the next section.

Results and Discussion
The Status of (T)àbí Re-Visited In the reviewed works above, scholars have claimed that (t)àbí functions as both a polar question and a conjunction.However, based on the earlier assertion that (t)àbi is a conjunction in all ramifications, the researcher presents the following arguments: i. Why is it that only (t)àbi functions as both a conjunction and a polar question whereas, no other conjunction can perform the same function and no other polar question word can act as a conjunction as well?ii.That polar question words are adjoined to a declarative sentence to form an interrogative sentence.The question is: how do we account for the instances where (t)àbi appears at the sentence initial position?iii.When (t)àbi co-occur with other polar question words, what specific function does it perform?If it is a conjunction, can it be a polar question as well?iv.Finally, why is it that it is only (t)àbi that can occur at the initial, medial and final position as a polar interrogative sentence while other polar question words are either restricted to the sentence initial, medial or final position?Based on the observations above, we present the following claims to support our arguments.
However, Polar question formation in Yorùbá involves the process of adjoining the polar question words to the existing declarative sentences.See the examples below: In the examples 9 (a-e), it is discovered that the polar question words are adjoined to the declarative sentence to form polar interrogative sentences.However, in examples (2 f and 7 b) if it is assumed that (t)àbí is adjoined to the declarative sentence as it appears at sentence initial position to produce the outcome like the ones in 9 above, let us consider instances where (t) A closer look at example 10 (a-b), will reveal that (t)àbí in these sentences is neither inserted nor adjoined.Bamgbose (1990, p. 190) observes among other things with respect to example (2 f and 7 b) and (10) that when two opposite sentences are combined with (t)àbi or where (t)àbi appears at a sentence initial position, it normally has interrogative connotation.He notes further that when (t)àbi occupies a sentence initial position, the clause before it has been deleted.However, this assertion in Bamgbose (1990)  In examples 11 (a-c), it is discovered that (t)àbí performs conjunction function as opposed to polar question as shown in 10 (a & b).These examples in 11 (a-c) point to the fact that the polar question words in 10 (a & b) have been deleted.Moreover, Ṣ é, ṣ èbí, njẹ ́ in (11 a-c) of the clause (S1) controls the structure in (S2).However, 11 b seems deviant; Adeoye (2015) argues that ṣ èbí is derived from ṣ é # àbí that is why the contiguous co-occurrence of ṣ èbí and àbí may not yield a grammatical sentence in Yorùbá.This assertion may look plausible but in Yorùbá a low tone cannot displace a high tone as shown in the concatenation.Moreover, this study still maintains that in the distant past in the historical development of Yorùbá sẹ ́ # àbí must have fused together to produce a polar question word because in some Yorùbá dialects the word is ṣ ebí; this is an indication that the merger of the high and low tone has produced a mid tone which is a plausible tonal process.This process is attested in Mọ bà dialect (see Abiodun & Ajiboye, 2008).It must be noted however, that ṣ èbí and ṣ ebí are in free variation in the Yorùbá language, that is, they have same semantic content.Another test to show that t(àbí) is not a polar question word is that no other polar question word such as ṣ é, ṣ èbí and njẹ ́ have the same distribution as t(àbí), that is, they must be able to conjoin two independent sentences.While examples in 12 (a-c) are ungrammatical, example 12 (d) is grammatical.This is a clear indication that (t)àbí is a conjunction.And, no real polar question word can combine two opposite sentences.Moreover, no other conjunction has been reported in Yorùbá to perform polar question word functions.If this is the case, why is (t)àbi an exception?However, with respect to the arguments above, three hypotheses are proposed by the researcher to account for (t)àbí in Yorùba.

Hypothesis (3) (t)àbí is a conjunction (disjunctive).
To start with the first hypothesis, the researcher proposes a Complementizer Phrase (CP) Structure for a polar question interrogative sentence where the polar question word occupies the C, CP position and it takes IP as its complement.Moreover, the study also proposes a ConjP that is dominated by IP x and it combines two sentences (IP n and IP m) together.The proposed structures are represented below: Based on our assertion in hypothesis ( 1), examples 14 (b, c and d) will have two reading as shown in 15 (b, c and d) diagrammatically above.However, example 15 d (i) shows that there are two sentences, Ṣ é questions the first sentence, while (t)àbi questions the second sentence whereas example 15 d (ii), depicts that there are two sentences; the first sentence is questioned by Ṣ é while (t)àbi combines the second sentence to polar interrogative sentence (first sentence).
The conclusion that will be drawn from the tree diagrams above is that the assertion in hypothesis (1) is ambiguous, because where (t)àbi is sandwiched between two sentences, it would either function as a polar question word or a conjunction.Moreover, when a polar question word is adjoined to the same sentence the function of (t)àbi will also remain the same.It must be noted that Yorùbá permits the co-occurrences of two polar question words in a sentence and these do not result in change of their functions.The examples in ( 16) above show that the occurrence of two or more polar question words in a sentence do not change the function of one of the two as opposed to example 15 d (ii) above.Moreover, it is pertinent to note that where (t)àbí co-occur with other conjunctions like yálà/bóyá its function remains as a conjunction.In addition, if t(àbí) is placed as a question word and a conjunction on the same node on a tree diagram, the structure will be cumbersome for analysis and it will generate a grammar that is hard to comprehend.
Since, Hypothesis (1) cannot provide a unified approach for the analysis of (t)àbí, it is pertinent to make recourse to hypothesis (2).In hypothesis (2), it is stated that (t)àbí is a polar question word in all its position of occurrence thus a CP structure is proposed where C projects into CP and it takes IP as its complement.The structures in 15 b (i) 15 c (i) and 15 d (i) will be applicable, but example 17 (a and b) will pose a problem to this assertion in hypothesis (2).See the tree diagram in 18 below:

18
The structure in 18 above and other examples where t(àbí) functions as a conjunction will show clearly that hypothesis (2) is not plausible.Thus, it is imperative to look for a unified analysis to capture t(abi) in all its position of occurrence in Yorùbá sentences.However, hypothesis (3) may offer a plausible explanation for the occurrences of t(àbí).
Hypothesis (3) asserts that all occurrences of (t)àbí is a conjunction.In line with this assertion, a ConjP-Structure is proposed, where (t)àbí occupies Conj that projects into ConjP and it takes IP or CP as its complement.Moreover, the ConjP can occur between IP n and IP m (declarative sentences).See the tree diagrams below: The assertion put forward in hypothesis (3) looks more plausible and it provides a uniform approach to the status of (t)àbí in all its places of occurrence.Based on this fact, it is argued that whenever (t)àbí occupies a sentence initial position, the clause or sentence before it has been deleted (see Bamgbose, 1990).It is also submitted that when it appears between two sentences, the polar question word has been deleted at the base.Thus, its (polar question word) presence becomes optional.The researcher also upholds the popular belief of Yorubá grammarians like Awobuluyi (1978) and Bamgbose (1990) on the use of voice raising to produce a polar question because of its cross linguistic evidence.This study asserts that voice raising to produce a polar question must have been responsible for the claim that t(àbí) is a polar question word by some Yorùbá scholars.It is also posited that there is no specific way to differentiate sentences with a raised voice and the sentences without a raised voice when they are reduced to writing.However, the researcher opines that Yes / No questions in Yorùbá can be interpreted just like the ɸ-features of N items based on its syntactic position (see Ilori, 2010).Thus, it has a weak [WH] feature that is why it can be deleted in some syntactic environment; and its presence may be optional in other environment.Finally, this paper submits that t(abi) is a conjunction in all its place occurrence in sentences.

Conclusion
This paper provides another argument that is quite different from the views of the earlier scholars.This paper challenges the traditional approach of the use of (t)àbí which ascribes two nomenclatures (lexical category -conjunction and functional category -polar question word) and use for it.The study establishes that the occurrences of t(àbí) in all syntactic environments is a case of a conjunction.The paper also asserts that where it appears at a sentence initial position, the question word in the sentence has been deleted.It is also established that when it is sandwiched between two sentences the question word has been deleted or its presence becomes optional.Moreover, this study submits that when it combines two sentences where the first clause is introduced by a polar question word, the same polar question word controls the second sentence but in utterances it is always deleted.However, this paper supports the claim that raising of the voice can be used to produce a polar question in Yorùbá and in such an utterance the polar question is always deleted.This research is limited to lexical categorization of t(àbí).Moreover, the deviance in the co-occurrence of ṣ èbí and àbí in an interrogative sentence needs further research, maybe an incursion into Yorùbá dialects may provide a clear answer.Furthermore, Yorubá lexical categories need a total reappraisal.In English language for instance, a definite number of lexical categories are taught in any grammar course.However, Yorubá grammar which has been undergoing research for decades has not evolved any definite number of word classes.This is a challenge that scholars cannot afford to run away from.
9 a.Adé lọ sí oko Declarative NP go to farm Ade went to the farm.b.Ṣ é Adé lọ sí oko?Qw NP go to farm Did Ade go to farm? c.Ṣ èbí Adé lọ sí oko?Qw NP go to farm Ade went to the farm, didn't he? d.Njẹ ́ Adé lọ sí oko?Qw NP go to farm Did Ade go to the farm?e. Adé ha lọ sí oko bí?NP Qw go to farm Qw Did Ade go to the farm?