Q UESTION N OUNS AND C LAUSE -T YPING I N Y ORÙBÁ

Interrogative sentences are deployed to perform speech acts of asking questions or making requests. Interestingly, the syntax of interrogatives in Yorùbá has attracted the attention of many researchers in the language (Ìlọ̀rí, 2010: Ọláńrewájú & Táíwò, 2020: Ọláńrewájú 2022). However, the correct feature specification and syntactic behaviour of question nouns (hence, QNs) are yet to be given adequate attention. Some of the extant works in the language equate QNs in Yorùbá with wh-phrases of English and some other Indo-European languages, This paper, within the confines of the Chomsky’s Minimalist Program, discusses the two existing views on how constituent interrogatives involving QNs are clause-typed (the traditional opinion hypothesizing that QNs are the question makers in constituent interrogatives while the other position holds the views that an abstract question morpheme clause-types a sentence as a constituent question). This paper lays some claims to support the traditional position that a clause is typed a constituent question by the transfer of the question force [QF] on a QN to the Inter0 through specifier and head agreement. Primary and secondary data were collected and subjected to syntactic analysis. QNs in the language are ta ‘who’ and kí ‘what’, èló ‘how much’, mélòó ‘how many’ and so on. Polysyllabic QNs are derivational in the language. Yorùbá content word questions, relative clauses, adverbials and so on are not signalled in wh-encripts unlike English. The language operates different functional heads to clause-type them.


INTRODUCTION
Interrogatives across word languages can be classified based on the types of responses they trigger into constituent and polar questions (König and Siemund, 2007;Issah, 2013).Unlike wh-phrases in English, QNs in Yorùbá are restricted to interrogatives only (Awóbùlúyì, 1978, 2008, 2013: Bámgbóṣé, 1990;Táiwò and Abimbọ ĺá, 2014;Ọláńrewájú, 2020, 2022 andso on).In Yorùbá, QNs, referred to as wh-phrases in some other literature, are onthologically different from their equivalents in English and some other European languages.Evidently, constituent questions are not signalled by wh-encripts in Yorùbá.Yorùbá operates different items (QNs, question verbs (hence, QVs) and interrogative qualifiers) to form its content word questions (Awóbùlúyì, 1978;Bámgbóṣé, 1990Bámgbóṣé, , Ọlańrewájú, 2022)).This paper classifies QNs and QVs as different lexical categories in Yorùbá, therefore, adopting wh-phrases for them would be descriptively inadequate and inappropriate.Ouhalla (1996) opines that wh-questions in natural languages differ in respect to their morphological and semantic properties.The veracity of this assertion is evident in the different ontological features of wh-phrases in English and QNs in Yorùbá as shown in the examples below: The English words above occur both as QNs and demonstrative adjectives and so on unlike their Yorùbá counterparts which are operated only as QNs.Let us also consider the English examples below: (2) a. i.Who did you see?
ii.The man who came here has left.Yoruba does not operate its QNs similarly to what is applicable in the types 2aii, bii and cii above because QNs in the language strictly occur in content word interrogatives.Saito (1992), in line with the the view above, claims that wh-feature exist in all languages.This is consequently responsible for the attraction of interrogative constituents to the clause left periphery, for a feature checking purpose.It is however discovered that the [+Q] feature on the Inter0 of a Yorùbá interrogative clause is not strong, therefore, it cannot trigger an overt movement of a QN (Ìlọ rí, 2010).Movement of an interrogative constituent to the clause left periphery is motivated by the strong [+foc] feature on the Foc0 (Ọláńrewájú and Táíwò, 2020).This prompted Issah (2013) to claim that interrogative constituents constitute a linguistic device used to identify pieces of information considered to be prominently new.Also, Kroeger (2004:139) in Issah (2013:56) opines that a question word bears pragmatic focus because it specifies the crucial piece of new information required.This paper has five sections: Section one discusses the introductory parts of the paper.Section two and three discuss the traditional position on Yorùbá QNs and the opinions of the recent scholarly works, that is, the opposing view respectively.In section four, the discussion is on features of Yorùbá QNs while the concluding remarks are drawn in Section five.

II
PREVIOUS STUDIES ON QNS IN YORÙBÁ Awóbùlúyì (1978Awóbùlúyì ( , 2013) )  Although, the QN mélòó (how many) begins with a consonant phoneme, it does not share this similar syntactic attribute with kí (what) and ta (who) due to its different structure and form of derivation.Melòó is derived from mú èló (pick/take how many) (Ọláńrewájú, 2022).This derivation is in line with Awóbùlúyì's (2008) claim on the derivation of Yorubá numerals like méjì, mẹ ́ta, mẹ ́rin and so on.These numerals do not also allow insertion of a mid-tone syllable (MTS) just like mélòó, as shown below: ( The examples in 5b and 6b are ill-formed.Suffice to note that the focused constituents in the pragmatic domain in 6 and 7 above lack semantic kinship unlike 4 and 5.This might reasonably factor why mélòó (how much) disallows MTS insertion unlike kí (what) and ta (who).
Yorùbá QNs according to Awobùlúyì (2013) can function as complements in Yorùbá clauses just like some other classes of nouns.In 8 below, èló (how much) functions as the complement of the verb dì (become).( 7) Ó di eló ba-yìí?
It become QN like-this 'How much is it now?(Awóbùlúyì, 2013:52) III FINDINGS AND DISCUSSION

ON THE OPPOSING VIEW
Ọláògún ( 2016), and Ọláògún and Aṣiwáju (2016) take a radical departure from the traditional position on QNs in Yorùbá by disregarding them as interrogative markers.Following Nkemnji (1995), and Aboh and Pfau (2011), Ọláògún and Aṣiwáju (2016) claim that items like ta, kí and so on only satisfy focus requirements and not clause-typing.They base their claim on the following evidence: i. Yorùbá operates an overt/abstract question morpheme to mark content word questions.ii.Wh-phrases co-occur with an overt/abstract question morpheme iii.Some other languages attest non-overt wh-phrase.iv.A wh-phrase does not mark questions alone in English.v. Yorùbá still operates wh-questions without wh-phrases.
The five points itemised above are subsumed under; (a) clause typing evidence, (b) information structure evidence and (c) clause structure evidence.
On the use of question morpheme stated (in i) above, Ọláògún (2016) claims that, just like some other languages under Kwa, Yorùbá operates an abstract question morpheme, and not QNs to clausetype a construction as content word question.According to him, the overt equivalent of the abstract question morpheme occurs after a subject DP, as evident (in 8a) below: Olu INTER know-emph 'Did Olu know?' (Ọláògún, 2016:14) The question morphemes a and ye come after the subject DPs and function as yes/no question markers in 8a and b.To Ọláògún, overt realisation of a question morpheme after the subject DP (in 8a) above is evidence that Yorùbá also operates its abstract equivalent, either after a subject DP or at the clause final position.It is however discovered that a is wrongly identified as yes/question marker (in 9a) above based on the following reasons:
Wise.person the PSM QV 'Where is the wise?' (1 Cor.1:20, Bíbélì Mímọ ) Example ( 9) above generates two plausible questions: one, considering a as a question morpheme in 10a above, how many question markers are operated in the derivation?Two, what clause-type the expression interrogative, that is, what is the question marker in the expression?The plausible answers are two and dà respectively.Dà is a content word question marker in Yorùbá (Awóbùlúyì, 1978(Awóbùlúyì, , 2013;;Bámgbóṣé, 1990 Another logical question generated by 9a and 10 repeated (as 14a and b) below, for ease of reference, is that if a, a question morpheme, according to Ọláògún (2016)  a functions as a pre-modifier in (13a and b) above.Therefore, Ọláògún (2016) still needs to adequately account for the true feature specification of a which he refers to as a question morpheme in 13a above.

OTHER COMMONLY USED YES/NO QUESTION MARKERS
The item a occurs with some other commonly used yes/no question markers like ǹjẹ ́ and ṣé as evident in the examples below: In a nutshell, ǹjẹ ́/ṣé is the yes/no marker (in 14a) above, and this indicates that, yes/no question marker is abstract in 10a repeated (as 13a) above.Similar to tilẹ , a functions as a pre-modifier (Táíwò, 2019;Ọláńrewájú and Táíwò, 2020).
On co-occurrence of wh-phrase with an overt or abstract question morpheme, Ọláògún (2016), and Ọláògún and Aṣiwájú (2016) also claim that evidence from other languages reveals that the equivalents of items like kí "what", and ta "who" in some other languages co-occur with overt question morpheme as shown (in 15) below: Now, the two germane questions that demand answers here are: What type of response does the abstract equivalent of yè trigger in Yoruba?Two, for the sake of intuition, how do we account for this question morpheme triggering two different types of responses: polar and constituent word answers?This implies that a is a pre-modifier just like tilẹ ̀ and lè (16b and c) below: To Nkemnji (1995), Aboh and Pfau (2011), Ọláògún (2016), and Ọláògún and Aṣiwájú (2016), the last two germane questions asked above are irrelevant.To them, focusing and clause-typing are teased apart; the question morpheme clause-types while QNs (identified as wh-phrases in their scholarly works) only satisfy focus requirement as shown in (18) phrase-marked as (19) below:  (Radford, 2009(Radford, : Ìlọ rí, 2010(Radford, : Ọlánrewájú, 2020(Radford, , 2022)) In 20a above, the QM wo performs interrogative function and it does not satisfy any focus reqiurement.Wo (an interrogative qualifier) and yẹn "that" (a qualifier) are not specified [+nominal], therefore, they cannot be hosted at the spec FocP.The DP aṣọ wo forms the question phrase (QP) (in 20a) above while the [+Q] feature wo percolates through the entire phrase aṣọ wo (Ajíbóyè, 2005).Also, extraction of the entire QP to the clause left periphery is in line with Wh-Attraction Condition (WAC) in (21) below: The edge feature on C attracts the smallest possible maximal project containing the closest whword to move to spec CP. (Radford, 2009:216) The implication borne out (20a and b) above is that once wo is the question marker in 20a, kí also marks question (in 22a) below: Ọlaògùn (2016) also observes that wh-phrases do not mark only questions in English, they also occur in declarative sentences, as shown (in 23 and 24) below.Consequently, the item is not marking constituent interrogatives in Yorùbá.A cursory look at 23a-c above reveals that who marks a constituent question in 24a, relativisation in 23a-c, and where as an adverbial marker in 24c.This consequently factors English adopting the "whterm".Yorùbá content word questions, relative clauses, adverbials and so on are not signalled in whencripts unlike English.Yorùbá operates different functional heads to project them.Let us consider the examples below for a better explanation: (25) a. [InterP Kí [Inter' ø [FocP <kí>[Foc' ni [TP Olú [vP <kí> [v' <Olú> [v' rí [VP <kí> Rel' ti [TP Olú [vP <ibi> [v'<Olú>[v' jókòó I see money at place REL Olú sit to 'I saw Olú where he sat.' A cursory look at the gloss in each of 25a-c above reveals that English operates a wh-expression in a wh-question in 25a, and also in relative constructions in 25b-c.This is not applicable in Yorùbá, where different lexical items are operated to mark content word questions and relative constructions.Therefore, QNs in Yorùbá are ontologically different from wh-phrases operated in English.
Ọláògún ( 2016) also claims that Yorùbá operates wh-questions without wh-phrases as shown below: 'Where is the book?' (Ọláògún, 2016:129) The data (in 26a-b) above suggest two facts: one, Yorùbá operates lexical items with inherent [+Q] feature to form its constituent interrogatives and examples are QNs, QVs and interrogative qualifiers.Two: dà and ńkọ ́ (in 26a-b) above are QVs, contrarily to Awóbùlúyì's (2013) position (Read Bámgbóṣé, 1990;Táíwò and Abímbọ ĺá, 2014;Ọláńrewájú, 2022).Therefore, examples (26a and b) above are sentences while dà and ńkọ ́ are their predicates.It is also equally important to note here that dà, ńkọ ́, ta, kí, wo and so on are used to form content word questions, therefore, they do not have the same categorial status.
Another plausible evidence that reveals how QNs do more than focus marking is shown in the examples from Central Yorùbá dialects as shown below: agreement.This implies that only the DP ibi 'place' and not kà (QN) undergoes focusing in (29) above.Kà is externally merged at the spec InterP in line with Radford's (2009: 124) proposal (29) below: (29) A clause is interpreted as a non-echoic question if (and only if) it is a CP with an interrogative specifier i.e a specifier with an interrogative word) (Ọláńrewájú, 2022: 167) 3.2 FEATURES OF YORÙBÁ QNS

DERIVATIONAL
Apart from kí (what) and ta (who) that are monosyllabic, other QNs have more than a syllable and they are all derivational.

COMPLEMENTS
They function as complements for (transitive) verbs and transitive prepositions according to Awóbùlúyì (2013) Interrogatives in 30a-c are echoic types, the QNs kí (what), mélòó (how many) and ibo (where) are legible to the PF interface at the canonical position associated with their grammatical functions.Under minimalist assumption, they take LF movement to the spec InterP (Radford, 2009).It is discovered that the QN for human referent, ta (who) does not occur in situ unlike some other QNs.This factors why the derivations (in 31a and b) below crash at the LF interface.

SUBJECT DP
A QN also functions as subject DP in a copula construction (Yusuf, 1990) In 32, the subject QN ta 'who' is copied to the spec FocP through the spec TP before it finally lands at the spec InterP.Ni as the copula assigns a nominative case to yin 'you' which implies that ni in 33 is not a focus maker here, it behaves like a verb (Adéwọlé, 1991).

QUALIFIER
QNs also function as qualifiers in Yorùbá, just like wo (which) qualifies a head noun.An interrogative qualifier is used with the head noun in 33a while QNs are used as qualifies in 33b-c.The head nouns inherit the QF in each of the constructions through percolation.Also, the question phrases (QPs) in the examples above are undetachable in line with Wh-Attraction Condition (WAC) discussed earlier in this paper (Ajíbóyè, 2006;Radford, 2009;Ìlọ rí, 2010).

RHETORICAL QUESTION
Two QNs can be stacked in rhetorical question forms in Yorùbá as shown below: The implication borne out of 34 above is that Yorùbá conflates QNs only in echoic interrogatives.Also, copying a QN to the clause left periphery is mainly determined by the actual QN a speaker intends to focus.Therefore, Attract the Closest Principle (ACP) captured under Superiority Condition in previous models of generative grammar is not observed in the language when QNs are stacked.

IV CONCLUSION
The syntax of interrogatives in Yorùbá is evidence that the language has some dissimilar features compared to English and some other Indo-European languages with respect to how it forms its questions.QNs are specified with [QF], therefore, they do more than satisfying focus requirements contrary to what Ọláògún (2016), Ọláògún and Aṣiwájú (2016) claim.Therefore, they are ontologically different to wh-phrases of English.In Yorùbá, constituent interrogatives and polar questions are clausetyped differently.Formation of a non-echoic question using a QN involves the syntactic movement of the QN to the clause left periphery whereby the Inter0 acquire the [QF] through specifier and head agreement in line with Cheng's (1991) Clause-Typing Hypothesis and Radford's (2009) proposals while a polar question is clause-typed by the [QF] on the overt or abstract question marker (the Inter 0 ).
b. i.What do you need?ii.I have seen what I needed.c. i.Where did you keep it?ii.I saw it where I kept it.
, the QN originates from the vP domain.The derivation goes thus: The lexical verb jẹ 'eat' merges with kí 'what' to project the V-bar jẹ kí 'eat what' in line with c-selection requirement of the verb.After this, the QN kí 'what' is copied to the spec VP by the Operation Copy and Delete to have its case feature checked through specifier and head agreement.The derivation proceeds by merging the null performative verb v0 with the VP to project the v-bar while the strong vF on the light v 0 attracts the lexical verb jẹ 'jẹ' to adjoin to itself.The second person singular subject pronoun o 'you' is externally merged as the inner specifier of the light verb phrase (vP) in line with the Predicate-Internal Subject Hypothesis (PISH).The QN kí is attracted to the outer spec vP, an escape hatch from Phase Impenetrability Condition (PIC).This invariably allows it visible to further operations in the course of the derivation.The derivation proceeds by externally merging the abstract T 0 with the vP to project the T-bar.The T 0 as a probe attracts the second person singular subject pronoun o 'you' to the spec TP to check its [+case, EPP] feature.The abstract Foc 0 merges with the TP to project the Foc-bar.The Foc 0 as a probe searches its c-command domain and attracts the QN ki 'what' (an active goal) to the spec FocP to have its [+Foc] feature checked.The derivation proceeds by externally merging the abstract Inter0 with the FocP to project the Inter-bar.The Inter 0 as a potential goal attracts the QN kí 'wdat' to the spec InterP to check its [+Q, EF].Only QNs can occupy the spec InterP in Yorùbá b above, the QN kí 'what' functions as interrogative qualifier while olùkọ ́ 'teacher' functions as nominal qualifier.
met the man whom you interviewed last week.b.The committee decided over who will represent the University at the meeting.c.The boy who bought a car last week is dead.Who broke the plate?b.I have seen the boy who broke the plate.c.I met the boy where he broke the plate.
triggers yes/no response in 13a, what type of response does it trigger in (11) repeated as (13b) below, if truly it is a question marker in Yorùbá?
. Another plausible fact revealing that QNs are inherently interrogative in Yorùbá is shown (in 20) below: , as shown below: