You are on page 1of 9

Agreement by default and V-movement in Igbo

Greg Obiamalu
Department of Linguistics
Nnamdi Azikiwe University, Awka
E-mail: gregobiamalu@yahoo.com

Abstract

It is the traditional assumption that Igbo has no morphologically realized agreement


marker. However, Dechaine (1993) observed that Igbo has a default agreement marker
which occurs only in negative constructions. This default agreement marker is the E-
prefix which was traditionally assumed to be part of a negative discontinuous morpheme
‘e…ghi’. The E- prefix is a default agreement marker because it occurs only with certain
kind of subject DPs. It occurs because V-movement to T is blocked by NEG and leaves T
empty thereby allowing a default agreement to surface. This paper seeks to provide
further arguments in support of the default agreement status of the E- prefix by looking at
other construction types in Igbo that also involve V-movement.

1.0 Introduction

Recently, the status of the so-called dependent weak pronouns: i/ị, o/ọ, and the
impersonal pronoun a/e has been a subject of controversy among some Igbo scholars.
Nwaozuzu (2007:25-26) using certain syntactic evidence such as: pronouns as second
objects, pronouns in relation to prepositions, etc, argues that the so-called impersonal
pronoun e/a is not actually a pronoun but a “remnant of the nominalizer or the nominal
marker associated with the Niger-Congo group of languages to which Igbo belongs”.
According to her, a sentence such as (1) containing such element should rather be
analysed as having an “empty category NP subject” which is standardly referred to as
‘small pro’ in the GB framework.

(1) pro Egbù-rù m̀madụ̀ ebe ahụ̀


pro kill-rV(past) place that
‘Somebody was killed there’

Anyanwu (2006: 41-48) provides further evidence in support of Nwaozuzu’s position but
only differs from her in arguing that the impersonal pronoun e/a behaves exactly the
same way as the second and third person singular pronouns i/ị and o/ọ when compared
with the substantive personal pronouns such as: ya, ha, anyi, unu. He rather calls them
Pronominal Subject Clitics (PSCs). Anyanwu (2006) also raises a very important issue
regarding the position of the PSCs in the clause structure. There are two possibilities.
One, PSC is in the Spec AgrP position and therefore an argument assigned a nominative
case. The second possibility is that the PSC occupies the head Agr position while the
Spec AgrP is occupied by pro and its phi-features (person and number) can be identified
from the PSC in the Agr head position. He, however favours the second analysis based on
certain syntactic evidence.
The present paper takes its premise from the Agreement Head analysis which
Anyanwu (2006) terms ‘the pro analysis’. We intend to provide further evidence in
support of the overt agreement marker in Igbo by looking at the status of the e-/a-
harmonising verbal prefix that usually occur in negative and some other constructions
where V-movement is involved iff the subject is a non-clitic. Dechaine (1993) refers to
the prefix as a default agreement marker. Before we go into the argument, let us discuss
some theoretical assumptions.

2.0 AGR as a functional category

Agr is an abbreviation for agreement. It was seen in the Government and Binding
Theoretic framework as one of the features of the head I(nfl). The head I of a finite clause
is said to have the features [+AGR, +TENSE]. Agreement is defined by Crystal (1985:11)
as “a traditional term used in grammatical theory and description to refer to a formal
relationship between elements, whereby a form of one word requires a corresponding
form of another to express person, gender, number, etc. In traditional grammar, the term
‘concord’ was also used to express this formal relationship between linguistic elements.
Pollock (1989) suggests that the INFL node which was analyzed in GB as having the
features Tense and Agr be split into two functional heads: Tense and Agr, each having its
own maximal projection: Tense Phrase and Agr Phrase respectively.
Within the Minimalist Program, the assumption is that the functional categories Agr
and T have agreement and tense features that may check off the corresponding features
(of agreement and tense) on a verb that moves up and adjoins to them. Furthermore, the
functional categories Agr and T also contains case and phi features (number, person,
gender, etc.) that they may check off against features of NP that raise to their Specs. The
basic assumption of the Minimalist Program is that “these functional nodes never contain
items from the lexicon, they are not the positions in which inflectional affixes are
inserted. Rather, inflectional affixes are attached to the items of the lexical category in the
lexicon. The functional nodes Agr and T serve only to carry the morphological
(inflectional) features of Ns andVs”. (Marantz 1995:363) This Implies that the verb for
example carries with it all the affixes from the lexicon and only moves to check off those
inflectional features with the relevant functional heads. This brings us to the discussion of
v-movement.

3.0 V-movement

The minimalist assumption is that the lexical items such as verbs, nouns and
adjectives are fully inflected in the lexicon. In other words, they are taken from the
lexicon fully inflected with inflectional affixes. The verb for example moves to check its
features against the features of the relevant functional heads. It is only by such
movement, that the verb can check-off its morphological features against the features of
the relevant functional heads. Failure to eliminate the morphological features which can
be visible at the interface levels: PF and LF, before spell-out causes the derivation to
crash. (cf. Marantz 1995, Chomsky 1995). This type of movement is known in the

2
literature as V-T movement. The verb moves to the head T position while the Subject DP
which is base generated in the Spec VP moves to the Spec AgrP position. This is
illustrated in (2).

(2)
AgrP

Spec Agr´
Agr TP
T VP
Spec V´

DPx Vi tx ti DP

4.0 Agr and T in Igbo

It has been a subject of controversy whether the functional category Tense exist in
Igbo. It has been claimed that ‘Aspect’ rather than ‘Tense’ is explicitly marked in Igbo.
(cf Emenanjo 1978, Manfredi 1991) We shall argue in this paper for the existence of both
Tense and Aspect as functional projections, as well as Neg and Agr in Igbo.
V-movement is attested for Igbo. Evidence for this lies with the fact Igbo verbs bear
different types of inflectional affixes such as tense, aspect, negation as well as agreement.
The following data in the table below show different verb forms with different
inflectional affixes

Past Tense form Perfective Aspect form Negative form

3a. Òbi gbù-rù agụ̄ 3b. Òbi e-gbū-o-la agụ̄ 3c. Òbi e-gbū-ghi agụ̄
Obi kill-rV (past) lion Obi Agr-kill-OVS-Perf lion Obi Agr-kill-Neg lion
‘Obi killed a lion’ ‘Obi has killed a lion’ ‘Obi did not kill a lion’
4a. Ha zà-rà ụlò̟ 4b. Ha à-zà-à-la ụlò̟ 4c Ha a-zà-ghi ụlò̟
3Pl. sweep-rV(past) house 3Pl. Agr-sweep-OVS-Perf house 3Pl. Agr-sweep-neg house
‘They swept the house’ ‘They have swept the house’ ‘They did not sweep the house’
5a. O rì-rì nri 5b. O ri-e-la nri 5c. Ò ri-ghi nri
3SgCl. eat-rV(past) food 3SgCl. eat-OVS-Perf food 3SgCl. eat-Neg food
‘He/She/It ate’ ‘He/She/It has eaten’ ‘He/She/It did not eat’

The presence of the rV suffix as the morphological marking of past tense as in (3a), (4a)
and (5a) is an indication that Tense and V-movement exist in Igbo.

3
(3b), (4b) and (5b) illustrate perfective aspect. Perfective aspect is marked by the -la
suffix in Standard Igbo. Uwalaka (2003:10) argues that since Igbo does not permit the co-
occurrence of tense and aspectual morpheme, then TP and AspP share the same head and
occur under one node. We shall argue against this assumption when we look at negative
constructions. A closer look at (5b) and (5c) reveals that when the subject is a clitic, the
verb does not take the harmonizing e-/a- prefix in perfective and negative constructions.
We shall henceforth refer to this prefix as E-

5.0 The status of the E- Prefix

5.1 The E-Prefix in Negative constructions

In negative constructions, the E- prefix occurs if the subject is a non-clitic. Whereas


in its affirmative counterpart, no such prefix occurs as shown in (6) – (9).

(6) a. Ezè rì-rì nri b. Ezè e-rī-ghī nri


eat-rV(past) food Agr-eat-Neg food
‘Eze ate food’ ‘Eze did not eat food’

(7) a. Ha mà-rà mmā b. Ha a-mā-ghī mmā


3P VR-rV(ass.) beauty 3P Agr-VR-Neg beauty
‘They are beautiful’ ‘They are not beautiful’

(8) a. O rì-rì nri b. Ò ri-ghī nri


3S eat-rV(past) food 3S eat-Neg food
‘He/She ate food’ ‘He/She did not eat food’

(9) a. I rì-rì nri b. I ri-ghī nri


2S eat-rV(past) food 2S eat-Neg food
‘You ate food’ ‘You did not eat food’

Why do non-clitic subjects such as Eze in (6) and Ha in (7) occur with the E- verbal
prefix in the negative construction but not in the affirmative? Emenanjo (1981), Clark
(1989) and Uwalaka (2003) argue that Igbo negation is marked by a discontinous
morpheme e…ghi. Uwalaka (2003:12) further argues that “the negative prefix is only
deleted when the subject NP is one of the singular pronouns in the language”. She gives
Economy Principle as the reason for such deletion. She does not however point out why
only singular pronoun should cause the deletion of the prefix. Dechaine (1993:135)
argues against the templatic analysis of Igbo negation. According to her, the E- prefix is
a default agreement marker which surfaces as a result of stranded Tense. “Neg between T
and V ́ is a barrier for V to T movement. As V can’t raise beyond Neg, T above Neg is
empty”. Since Tense is not morphologically realized in negative constructions, the
agreement morpheme has to surface. The clitic pronouns do not co-occur with the
agreement prefix because they occupy the same Agr head position. This is line with

4
Anyanwu’s (2006) analysis that the PSCs occupy the Agr head while the Spec AgrP is
occupied by a null DP pro. (7b) and (8b) are shown on the tree diagram below

(10) AgrP
Spec Agr´
Agr TP
Spec T´
T NegP
Neg VP
Spec V´
V DP

7b. Hax a- mā-ghīs tx ts mmā


8b. prox Ò ri-ghīs tx ts nri

5.2 E- prefix in Perfective construction

The presence of the E- prefix in perfective constructions where no negative


meaning is implied supports the rejection of templatic analysis of Igbo negation as E…
ghi. Consider (11) below

(11) a. Àda a/à-zà-a-la ụlò̟ c. Ọ zà-a-la ụlò̟


Agr-sweep-OVsuff-Perf house 3S sweep-OVsuff-Perf house
‘Ada has swept the house’ ‘He/She has swept the house’

b. Ha e/è-gbu-o-la agū d. E gbu-o-la m̄ agū


3P Agr-kill-OVsuff-Perf lion Agr kill-OVsuff-Perf 1S lion
‘They have killed a lion’ ‘I have killed a lion’

If the same Agr prefix occurs in perfective constructions, how do we reconcile it with the
Neg barrier analysis proposed by Dechaine (1993)? Obiamalu (2006:28) argues that Igbo
has AspP that occupies a distinct position from TP. This is contrary to Uwalaka’s
(2003:10) assumption that since Igbo does not permit the co-occurrence of an aspectual
morpheme with tense morpheme that TP and AspP share the same head and occur under
one node. V moves to Asp and since the aspectual morpheme and the tense do no co-
occur in Igbo, the complex structure V+Asp does not move further to T. This leaves T
stranded just like in negation, thereby triggering of the emergency of the default
agreement prefix. Tense has been shown to be a universal functional category whether it

5
is morphologically marked or not. An aspectual construction encodes Tense meaning
even though the tense morpheme may not be morphologically realized as in Igbo.

5.3 E- prefix in causative construction

Anyanwu (2006) and (2007) discuss extensively the syntax and semantics of Igbo
causative constructions. He observes that a syntactic causative construction is bi-clausal
where the causative verb takes an AGRP (clausal) complement. He further notes that the
subject of the complement clause could either be a substantive NP, an accusative pronoun
or a Pronominal Subject Clitic (PSC). The following examples are adapted from
Anyanwu (2006:50)

(12) a. Ezè mè-rè ọ ta-a anụ̄


Eze cause/make-rV(past) 3sg.CL chew-OVSuff. Meat
‘Eze caused/made him/her eat meat’

b. Ezè mè-rè ya à-ta-a anụ̄


Eze cause/make-rV(past) 3sg.Pron Agr-chew-OVSuff. Meat
‘Eze caused/made him/her eat meat’

c. Ezè mè-rè Àda à-ta-a anụ̄


Eze cause/make-rV(past) Ada Agr-chew-OVSuff. Meat
‘Eze caused/made Ada eat meat’

While Anyanwu (2006) rightly observes that the PSC ọ and the substantive pronoun ya
are formally distinct, we differ from him in saying that the two do not occur “in identical
environment”, that is, do not occupy the same position as he claims. He seems to have
ignored the presence of the verbal prefix that occurs only with the non-clitics as in (12b)
and (12c) . While we agree with him that the PSCs are base generated under Agr, we
differ from him in analyzing the E- prefix occurring only with the non-PSCs as equally
occupying the Agr head position. The E- prefix in the causative construction, just like in
negative and perfective, isAgrP1
a default agreement marker.
Specand (12b)
The structure of (12a) Agr´are shown on the tree diagram in (13) below

Agr TP
T VP1
Spec V´
V AgrP2
Spec Agr´
Agr TP
(13)
T VP2
Spec V´
V DP
6
12a. Ezèx mèrèi tx ti pros ọ ts taa anụ̄
12b. Ezèx mèrèi tx ti yas à- ts taa anū
If the E- prefix on the verb of the complement clause of the causative construction is an
agreement marker, how does it relate to the functional category barrier analysis earlier
proposed for negative and perfective constructions where such prefix also occurs.
Our proposal here is that the verb of a complement clause of the causative
construction does not undergo v-movement. Since there is no V-movement, T remains
empty and the default agreement prefix emerges at the Agr head position. This is similar
to Igbo serial verb construction like (14) where only the first verb in the series bears the
tense suffix and moves up to T to check off its tense features. The subsequent verbs bear
no tense affix and therefore remain in-situ.

(14) Àda gbù-rù ọ̀kùkọ̀, bọ̀-ọ, sì-e, rì-e


Ada kill-rV(past) fowl dissect-OVS cook-OVS eat-OVS
Ada killed a fowl, dissected it, cooked and ate it

The subsequent verbs: bọ, si and ri bear no inflectional affix. The only element attached
to the verb root is the open vowel suffix which carries no meaning. It may have been
attached to the verb because Igbo verb roots do not occur alone. The OVS could be found
in some other constructions such as perfective (as in (11), imperative, etc., without any
identifiable meaning.
The structure in (15) is ill-formed because V-movement occurred in the complement
clause.

(15) * Ezè mè-rè Àda tà-rà anụ̄


Eze cause/make-rV(past) Ada chew-rV(past) meat

7
It is interesting to ask why the second verb in a causative construction does not need to
move to T. Anyanwu (2007:93-94) provides a clue. According to him, “the verb of the
complement clause necessarily refers to a sub-part or aspect of a single over-all event.
For this reason, the embedded verb acquires covertly the tense properties of the matrix
verb. This means that tense, having been overtly marked on the matrix verb becomes a
feature of the embedded verb; hence, it is redundant to mark it again”. (15) violates the
economy principle because the V-movement is ill-motivated. If this assumption is true it
then means that the TP projection of the matrix clause has scope over the embedded
clause and therefore carries the tense features of the embedded clause. The verb of the
complement clause need not move in order to check off its tense features.

6.0 Conclusion

We have tried to argue in this paper that that Igbo has overt morphological agreement
marker. This agreement marker occurs in certain syntactic constructions: negation,
perfective and complement clause of causative construction. The agreement marker is in
form of ‘Inflection by default’ (cf Manfredi 1993) because it occurs when V-T movement
fails to occur. T remains empty and would have been stranded if an overt morphological
element in form of agreement morpheme is not inserted. Dechaine (1993) likens this to
the English do-support, in the sense that E- gives morphological content to an otherwise
stranded T position. The E- agreement prefix does not co-occur with the pronominal
subject clitic because the clitic also occurs under the Agr head position. The overall
implication of this analysis is that the presence or absence of the agreement morpheme in
Igbo is determined by whether V-movement to T occurs at all or block along the way by
some other functional heads.

References

Anyanwu, Ogbonna N. 2006. Pronominal clitics in Igbo. Journal of the Linguistic


Association of Nigeria (JOLAN) No 9. 39-55.

Anyanwu, Ogbonna N. 2007. The syntax of Igbo causatives: A minimalist account.


Landmark Series 2. Port Harcourt: M&J Grand Orbit Communications Ltd.

Chomsky, Noam. 1995. The minimalist program. Cambridge, MA: MIT Press.

Clark, Mary. 1989. The tonal system of Igbo. Foris, Dordrecht

Crystal, David. 1985. A dictionary of linguistics and phonetics 2nd ed. Oxford: Blackwell.

Dechaine, Rose-Marie. 1993. Negation in Igbo and Yoruba. In Niger Congo syntax and
semantics 6. ed. by V. Manfredi, 135-150

Emenanjo, Nolue. 1978. Elements of modern Igbo grammar. Ibadan: University Press.

8
Emenanjo, Nolue. 1981. Auxiliaries in Igbo syntax. University of Ibadan Ph.D
dissertation

Manfredi, Victor. 1991. Agbọ̀ and Ehugbọ̀ : Igbo linguistic consciousness, its origin and
limits. Harvard University dissertation.

Manfredi, Victor. 1993. Inflection by default. In Niger Congo syntax and semantics 6, ed.
by V. Manfredi, 91-112.

Marantz, Alec. 1995. The Minimalist Program. In Government and binding theory and
the minimalist program, ed. by Gert Webelhuth, 349-382. Cambridge, MA:
Blackwell.

Nwaozuzu, G.I. 2007. Erroneous assumptions in the description of some aspects of


indigenous Nigerian languages: The case of the unspecified pronominal element in
Igbo. In Language and Literature in a developing country: Essays in honour of Prof
B.O.A. Oluikpe. ed. by B.N. Anasiudu, G.I. Nwaozuzu and C.N. Okebalama, 15-27.
Onitsha: Africana-First Publishers

Obiamalu, Greg O. 2006. The morphosyntactic spell-out of the functional category neg in
Igbo. Awka Journal of Linguistics and Languages (AJILL). Vol 2: 25-36.

Pollock, Jean-Yves. 1989. Verb movement, universal grammar and the structure of IP.
Linguistic Inquiry, 20: 365-424

Uwalaka, Mary A. 2003. Tense and v-movement in Igbo. Paper presented at the 4th
World Congress of African Linguistics. Rutgers University, New Brunswick, NJ.
17-22 June

You might also like