A theoretical and descriptive examination of Yorùbá discontinuous lexical predicates, their causative and inchoative behavior, and the challenge they present to the Lexical Integrity Hypothesis.
Yorùbá splitting verbs are composite verbal lexemes whose constituent morphemes obligatorily separate in transitive clauses to surround their direct object complement [S1, S2]. While canonical Yorùbá verbs are monosyllabic roots conforming to a Consonant-Vowel (CV) template, splitting verbs function semantically as unified, often non-compositional predicative concepts despite surfacing as two discontinuous words in syntax [S1, S3]. This phenomenon creates a fundamental tension for linguistic theory because the standard Lexical Integrity Hypothesis asserts that syntactic operations cannot access, interrupt, or manipulate the internal structure of words formed in the lexicon [S4, S5].
This file provides an exhaustive account of the morphology, syntax, and semantics of Yorùbá splitting verbs. It examines their core diagnostics, analyzes the major theoretical debates between lexicalist, serialist, and minimalist frameworks, details their behavior across causative-inchoative alternations and aspectual modifications, and situates their documentation in the historical evolution of Yorùbá linguistic study.
Standard Yorùbá is an isolating language exhibiting strict Subject-Verb-Object (SVO) baseline clause structure [S1, S6]. In standard transitive clauses, a simplex lexical verb directly precedes its nominal or pronominal complement without morphological disruption [S1, S6].
In splitting verbs (ọ̀rọ̀-ìṣe aṣèpín, literally "verbs that cause division", from ọ̀rọ̀-ìṣe, verb, and aṣèpín, that which divides), a disyllabic verbal item splits into two distinct segments, traditionally labelled V1 and V2, whenever a direct object is present [S1, S2]. The resulting linear surface order is Subject-V1-Object-V2 [S2, S7].
Canonical SVO Clause:
Adé jẹ ẹran náà.
Adé eat meat the
"Adé ate the meat."
Splitting Verb Clause (bàjẹ́, "to spoil/ruin"):
Adé ba iṣu náà jẹ́.
Adé V1 yam the V2
"Adé spoilt the yam."
If the two halves of a splitting verb remain adjacent before a direct object, the resulting structure is categorically ungrammatical in standard Yorùbá: *Adé bàjẹ́ iṣu náà cannot surface under any pragmatic condition [S2, S7]. Conversely, in intransitive contexts, or in clauses where the verb undergoes an inchoative alternation without an internal direct object, the two halves typically appear string-adjacent as a single compound verb: Iṣu náà bàjẹ́ ("The yam spoiled") [S1, S7].
The central descriptive difficulty of splitting verbs lies in their variable degrees of semantic compositionality and root autonomy [S2, S3]. In some splitting verbs, both V1 and V2 survive in contemporary Yorùbá as independent lexical items:
In other splitting verbs, one or both components have lost their independent syntactic distribution and semantic autonomy, existing exclusively within the discontinuous frame [S1, S2]:
The Lexical Integrity Hypothesis (LIH), formulated across various versions by Noam Chomsky, Mark Aronoff, Anna Maria Di Sciullo, Edwin Williams, and Joan Bresnan, posits that syntactic rules neither manipulate nor have access to the internal structure of words [S4, S5, S9]. Under strict lexicalist assumptions, morphological word formation takes place entirely within an autonomous lexicon prior to syntactic insertion; once a word is generated as an X0 head, syntax treats it as an indivisible, atomic unit [S4, S5].
Yorùbá splitting verbs present an acute challenge to this principle because they display properties of single words and multi-word syntactic phrases simultaneously [S2, S5, S7]:
This duality forced Africanist and theoretical linguists into long-running debates over whether Yorùbá splitting verbs prove that lexical integrity is false, or whether splitting verbs are not words at all but syntactic configurations derived by independent phrase-structure principles [S2, S5, S8, S11].
Scholarship on Yorùbá splitting verbs over the last six decades divides into five primary theoretical models.
+------------------------------------+----------------------------+------------------------------------------------------+
| Theoretical Model | Key Proponents | Core Analytical Claim |
+------------------------------------+----------------------------+------------------------------------------------------+
| Serial Verb Reductionism | Ayọ̀ Bámgbóṣé | Splitting verbs are underlying serial verb chains; |
| | | non-compositionality is idiomatic lexical storage. |
+------------------------------------+----------------------------+------------------------------------------------------+
| Inherent Discontinuous Lexeme | Oládélé Awóbùlúyì | Splitting verbs are distinct discontinuous lexical |
| | | primitives that undergo sandwich complementation. |
+------------------------------------+----------------------------+------------------------------------------------------+
| Semantic Extension / Root Decomp. | Yíwọlá Awóyalé | Component roots retain archaic semantic primitives; |
| | | all splitting verbs are fully compositional at base. |
+------------------------------------+----------------------------+------------------------------------------------------+
| Phrasal Morphology | Douglas Pulleyblank, | Word-formation rules operate post-syntactically; |
| | Akinbiyi Akinlabi | syntax feeds morphology, violating rigid LIH. |
+------------------------------------+----------------------------+------------------------------------------------------+
| Minimalist Head-Movement | O. G. Bode, Alicia Parrish,| Complex lexemes merge low in VP; V1 moves to |
| | Cara Feldscher, F. Ìlọ̀rí | functional light verb (v) or Asp, stranding V2. |
+------------------------------------+----------------------------+------------------------------------------------------+
Ayọ̀ Bámgbóṣé argued that splitting verbs are not unique morphological primitives, but rather a subset of Serial Verb Constructions (SVCs) [S8, S11, S12]. In Bámgbóṣé's analysis, Yorùbá syntax allows sequences of verbs sharing an object in the linear configuration NP-Subj + V1 + NP-Obj + V2 [S8, S12].
For Bámgbóṣé, a sentence like Ó ba aṣọ jẹ́ ("He ruined the cloth") shares the identical syntactic deep structure of an ordinary modifying or resultative serial construction like Ó mú aṣọ wá ("He brought the cloth", literally "He took cloth came") [S8, S12]. The fact that ba...jẹ́ has a unitary lexical meaning is treated as a matter of semantic specialization and idiomatic crystallization rather than a distinct syntactic category [S8, S12].
Opponents of Bámgbóṣé's model point out that genuine serial verbs allow each verb to be independently negated, focused, or modified in many contexts, whereas splitting verbs act as an indissoluble predicate: one cannot focus jẹ́ alone or negate ba without negating the entire event [S2, S7].
Oládélé Awóbùlúyì strongly rejected the reduction of splitting verbs to serial verbs, demonstrating that splitting verbs possess syntactic and distributional constraints that separate them entirely from SVCs [S2, S13].
Awóbùlúyì established that splitting verbs constitute an autonomous subcategory of compound verbs that undergo sandwich complementation [S1, S2]. Under this view:
Awóbùlúyì maintained that treating these verbs as ordinary serial verbs obscures their unique status as single words whose syntactic realization happens to be non-contiguous [S2, S13].
Yíwọlá Awóyalé argued against the claim that splitting verbs are semantically opaque or arbitrary . Awóyalé proposed that even seemingly opaque splitting verbs like bàjẹ́ or réjẹ ("to cheat, swindle") can be decomposed into primitive semantic roots when analyzed through cognitive semantics and historical Yorùboid root phonology [S3, S14].
Under Awóyalé's framework, V1 represents an initial action, contact, or stance (e.g., bà, to come into physical contact with, touch, or affect), while V2 specifies the resultant state or terminal consequence (e.g., jẹ́, to consume, extinguish, or alter state) [S3, S14]. Awóyalé argued that what modern speakers perceive as an unanalyzable idiom is the result of semantic generalization over highly productive Proto-Yorùbá resultative compounding .
However, critics like Awóbùlúyì and later generative syntax researchers pointed out that Awóyalé's semantic assignments often require attributing abstract meanings to bound morphemes that cannot be verified in any synchronic syntactic frame [S2, S7].
In their influential 1988 study, Douglas Pulleyblank and Akinbiyi Akinlabi utilized Yorùbá data, including splitting verbs and nominalizations, to challenge the standard generative separation between syntax and morphology .
Pulleyblank and Akinlabi demonstrated that Yorùbá exhibits phrasal morphology, where morphological processes such as prefixation and vowel deletion operate on syntactic phrases (XPs) rather than strictly on lexical roots (X0) . In nominalizations derived from splitting verbs (such as ìbànújẹ́, "sadness", derived from ì- + bà inú jẹ́), the morphological prefix ì- attaches to a syntactic verb phrase containing an internal nominal argument (inú, "inside/stomach") [S5, S7].
Because the morphological derivation encompasses a phrasal constituent that has already undergone syntactic complementation, Pulleyblank and Akinlabi concluded that the strict Lexical Integrity Hypothesis is untenable for Yorùbá: the syntax must directly feed morphological word construction .
In modern generative linguistics, researchers have accounted for splitting verbs using VP-shell architectures and Minimalist head movement [S7, S15, S16].
Alicia Parrish and Cara Feldscher expanded this analysis by demonstrating that word-order shifts triggered by the aspectual marker tún provide direct empirical proof of this movement path .
Minimalist Derivation Schema (Transitive bàjẹ́):
[vP Adé [v' ba_i [AspP [DP iṣu náà] [Asp' [VP t_i jẹ́ ]]]]]
│ │
└──────── Raised from V ───────┘
Splitting verbs exhibit several rigorous morphosyntactic diagnostics that differentiate them from other verbal categories in Yorùbá grammar.
A substantial subclass of splitting verbs undergoes a causative-inchoative alternation [S1, S7, S16]. In causative clauses featuring an overt agentive subject, the verb obligatorily splits around the theme object; in inchoative clauses where the theme argument surfaces as the grammatical subject, the two verbal pieces remain adjacent [S1, S7, S16]:
Causative (Split transitive):
Olú ba àga Òjó jẹ́.
Olú V1 chair Òjó V2
"Olú broke Òjó's chair."
Inchoative (Unsplit intransitive):
Àga Òjó bàjẹ́.
chair Òjó spoil
"Òjó's chair got spoiled / broke."
In the inchoative construction, because no external agent is projected into the clause, the functional light verb v does not induce V1 raising, allowing the full complex verb to surface as a unified phonetic unit [S7, S16].
The behavior of the repetitive and additive aspectual marker tún serves as one of the most powerful diagnostics for the syntactic architecture of splitting verbs [S7, S17].
When tún expresses an additive meaning ("also"), it attaches high in the clause as an adverbial adjunct above the verb phrase, leaving the standard S-V1-O-V2 order undisturbed [S7, S17]:
Additive "also" reading (S-V1-O-V2 preserved):
Adé tún ba iṣu náà jẹ́.
Adé also V1 yam the V2
"Adé also ruined the yam."
However, when tún expresses a repetitive aspectual meaning ("again"), it occupies a low functional aspectual head (Asp0) between the light verb v and the lexical VP [S7, S17]. This low attachment blocks the head movement of V1 to v. Consequently, the direct object is shifted to the left while both halves of the splitting verb remain adjacent on the right, yielding a surface Subject-Aspect-Object-V1V2 order [S7, S17]:
Repetitive "again" reading (SOV order with unsplit verb):
Adé tún iṣu náà bàjẹ́.
Adé again yam the ruin
"Adé ruined the yam again."
This diagnostic, documented by Parrish and Feldscher, establishes that splitting is entirely a syntactic derivation driven by verb movement rather than a static property of lexical storage .
When the complement of a splitting verb is a third-person singular pronominal clitic (a/e/ẹ/o/ọ/u/i, harmonizing with the preceding vowel or surfacing as a lengthened vowel), the pronoun is cliticized directly onto V1, sitting between V1 and V2 [S1, S2]:
Ó bà á jẹ́.
3SG V1 3SG V2
"He/She ruined it."
*Ó bàjẹ́ ẹ. (Ungrammatical)
In standard Yorùbá phonology, the high-tone pronominal clitic -á attaches to the monosyllabic verb bà, shifting the tone of the verb from low to mid by regular tonal sandhi, producing [ó ba á jẹ́] [S1, S8]. The internal pronoun is fully governed by V1 for accusative case licensing [S5, S16].
When splitting verbs are converted into abstract nominals, agentive nouns, or gerunds, they display strict morphological fusion, demonstrating their underlying status as single lexical targets [S1, S5, S18]:
In every nominalization, the nominalizing prefix binds to the left edge of the unified compound, and internal object arguments become permanently trapped within the compound structure, completely inaccessible to outside syntactic modification [S5, S18].
The following section presents core splitting verbs in standard Yorùbá, analyzed through the required philological and translation layers.
Original:
Olóṣèlú yẹn ba orúkọ ìdílé rẹ̀ jẹ́.
Literal Gloss:
Olóṣèlú yẹn ba orúkọ ìdílé rẹ̀ jẹ́.
Politician that strike name lineage his/her consume/ruin
Idiomatic English:
"That politician tarnished the reputation of his family."
(Translated by the corpus authors, confidence: high)
Original:
Olùkọ́ fi ìwé tuntun náà hàn àwọn akẹ́kọ̀ọ́.
Literal Gloss:
Olùkọ́ fi ìwé tuntun náà hàn àwọn akẹ́kọ̀ọ́.
Teacher take book new the appear PL student
Idiomatic English:
"The teacher showed the new book to the students."
(Translated by the corpus authors, confidence: high)
Original:
Àṣà tí a ti dìmú láti ìran dé ìran.
Literal Gloss:
Àṣà tí a ti dì-mú láti ìran dé ìran.
Custom that 1PL PRF bind-catch from generation reach generation
Idiomatic English:
"The tradition that we have maintained from generation to generation."
(Translated by the corpus authors, confidence: high)
Original:
A kì í ba tẹni jẹ́,
kí á tó fi tòde bọ̀ ọ́.
Literal Gloss:
1PL NEG HAB strike POSS.own ruin,
that 1PL CMPL take outside replace 3SG.OBJ
Idiomatic English:
"One does not ruin one's own affairs,
only to replace them with what belongs to the public outside."
(Collected and translated by Oyekan Owomoyela, 2005)
The documentation and theoretical understanding of Yorùbá splitting verbs has evolved across two centuries of linguistic history, tied to changes in orthography, grammatical theory, and language contact.
The earliest written accounts of Yorùbá verbal morphosyntax emerged from liberated Yorùbá returnees and European Christian missionaries in Freetown, Sierra Leone, and Abeokuta [S22, S23].
In his landmark Grammar and Vocabulary of the Yoruba Language (1852), Bishop Samuel Ajayi Crowther noticed that specific verbs required what he described as a split placement around direct nouns, recording forms such as bà...jẹ́ and fí...hàn . However, Crowther lacked a theoretical framework for discontinuous constituents and treated them primarily as compound idioms where particles were inserted between words .
Reverend Thomas Jefferson Bowen, in his 1858 Grammar and Dictionary of the Yoruba Language, attempted to classify these items through classical Latin grammatical categories, misidentifying V1 as an inseparable preposition and V2 as a transitive verb . Bowen's classification reflected nineteenth-century European missionary biases, which struggled to interpret isolating Niger-Congo morphosyntax without forcing it into Indo-European inflectional paradigms .
During the height of the Ọ̀yọ́ Empire (seventeenth to late eighteenth centuries), court speech, royal bards (akéwì), and oral historians (onírànkín) utilized splitting verbs as highly flexible rhetorical instruments in praise poetry (oríkì) [S19, S24].
In central and eastern Yorùbá dialects (such as Èkìtì, Ìjẹ̀ṣà, Ìkálẹ̀, and Òwọ̀), splitting verbs exhibited phonetic forms distinct from the standard dialect that later coalesced around Ọ̀yọ́-Ìbàdàn speech [S16, S25]. For example, eastern dialects frequently preserved older unmerged vocalic sequences in the second verbal element, such as bà...lẹ́rù jẹ́ or dú...mú [S16, S25]. The nineteenth-century civil wars (Yorùbá Wars, 1820s-1893) displaced millions of speakers into major refugee camps and military strongholds such as Ìbàdàn, Abẹ́òkúta, and Ìjàyè, accelerating the dialectal leveling that cemented the Ọ̀yọ́-based Standard Yorùbá splitting paradigms .
During the British colonial administration of Nigeria, language manuals designed for colonial officers and educators began documenting splitting verbs with increasing descriptive precision [S26, S27].
Ida C. Ward, in An Introduction to the Yoruba Language (1952), recognized the rigid tonal patterns of discontinuous verbs, providing the first precise phonetic and tonal transcriptions showing that the tone of V1 systematically alternates based on whether its complement is a noun or a clitic pronoun . E. C. Rowlands later expanded this work in Teach Yourself Yoruba (1969), terming them "separable verbs" and emphasizing to learners that the split is obligatory in transitive constructions .
The formalization of African linguistics following Nigerian independence in 1960 turned Yorùbá splitting verbs into a major battleground for general grammatical theory [S2, S5, S8, S12]:
In the twenty-first century, research on Yorùbá splitting verbs has focused on two fronts: formal generative syntax and diaspora contact linguistics [S7, S16, S29].
*Wọ́n spoil project náà jẹ́ is ungrammatical) . The morphosyntactic frame of the splitting verb remains strictly governed by the Yorùbá matrix grammar .In Afro-Atlantic religious traditions, particularly Lucumí in Cuba and Candomblé Ketu in Brazil, liturgical songs preserve fossilized splitting verbs in sacred chants, though without active syntactic productivity among non-fluent practitioners [S10, S24].
The systematic documentation, analysis, and preservation of Yorùbá splitting verbs has been carried forward by generations of native and international scholars, writers, and institutions:
Despite extensive research, several structural questions regarding Yorùbá splitting verbs remain subjects of live investigation:
A scholarly analysis of major reference grammars, foundational syntactic debates, and dialectological classifications in Yorùbá linguistics.
The machinery Yorùbá uses to make new words, and a worked, confidence-flagged examination of the etymologies of the key religious and philosophical terms.
A comprehensive linguistic investigation into the Yorùbá High Tone Syllable (HTS), detailing its phonological behavior, syntactic distribution, dialectal variation, and competing theoretical models across six decades of scholarship.
An examination of the syntax, headedness debates, and argument sharing mechanisms of Yorùbá serial verb constructions across generative, typological, and historical frameworks.
An analysis of Yorùbá information packaging, examining focus fronting, content wh-interrogatives, the syntactic status of the marker ni, predicate clefting, and subject extraction asymmetries.
What Yorùbá is as a language, which languages it is related to, how its dialects vary across Yorùbáland, and how many people speak it today.