This paper reports a case of NOM/ACC alignment in agreement that results from the relative timing of A-movement and ф-agreement, and not from the position of ф-probes or the sensitivity of agreement to case. The NOM/ACC phenomenon discussed here is object agreement in Ndebele, which can be characterized as the systematic inability of Voice to agree with the highest argument, despite c-commanding it. I argue that this pattern follows from a general property of Bantu languages whereby every ф-probe obligatorily co-occurs with an
Bantu languages are characterized as having a NOM/ACC alignment in agreement. In many of them, the specific agreement pattern is the following: one ф-probe agrees with the highest DPs in the clause (“subject agreement”, italicized in (1)), and another ф-probe agrees with a lower DP (“object agreement”, bolded in (1)).
(1)
1s-
‘Thabani cooked the meat.’
Despite being well known, this pattern in Bantu languages is not well understood. As I argue here, it cannot be derived from fine-tuning the positions of ф-probes in the clause or from case-discrimination in agreement. Focusing on Zimbabwean Ndebele, I propose that the NOM/ACC agreement pattern can be deduced from two existing hypotheses. The first hypothesis is that ф and
In section 2, I describe the agreement pattern in Ndebele. There are two ф-probes in the Ndebele clause: one in T and one in Voice. Importantly, both probes c-command
(2)
The NOM/ACC phenomenon analyzed in this paper
The pattern in (2) is manifested in the inability of the thematically highest argument to control object agreement – whether the highest argument moves to Spec,TP (3-a) or stays in situ (3-b).
(3)
a.
w-a-(
1s-
i-nyama.
‘Thabani cooked meat.’
b.
Kw-a-(
15s-
i-nyama.
‘Thabani cooked meat.’
In section 3, I consider three possible sources for the NOM/ACC agreement pattern in Ndebele: i) fine-tuning the position of the object agreement probe, ii) case-discriminating agreement and iii) countercyclic probing by T before Voice. I then demonstrate that independently known properties of Bantu languages make it very difficult to pursue the first two types of analyses, and conclude that the countercyclic analysis is undesirable on theoretical grounds.
Section 4 develops the proposal, capitalizing on an existing hypothesis according to which ф-probes in Bantu languages obligatorily cooccur with
(4)
The ф–
In section 5, I discuss the implications of the proposed analysis for the connection between object agreement and object dislocation. Even though agreed-with objects must be dislocated in Ndebele, the object agreement probe and the object-dislocation probe are not located on the same head. The reason why the two operations cooccur is that the triggering probes (
Section 6 further supports the proposed analysis by demonstrating that it derives three seemingly unrelated properties of Ndebele ф-agreement and A-movement: defective intervention effects in VSO clauses (6.1), the optionality of subject raising and agreement (6.2), and an object agreement asymmetry in passives of ditransitives (6.3). Section 7 concludes the paper and situates its findings in a crosslinguistic context.
The core data in this paper come from Zimbabwean Ndebele (Bantu, Nguni group, S44), but many of the phenomena discussed here are very robust across the Bantu language family and have been discussed in the literature for related languages, such as Zulu (i.a.
Ndebele exhibits subject agreement and object agreement (5). Subject agreement is obligatory, while object agreement appears only when the object is discourse-given.
(5)
a.
U-Thabani
1s-
i-nyama.
‘Thabani will cook the meat.
b.
U-Thabani
1s-
i-nyam-a.
‘Thabani will cook meat.’
The object marker in southern Bantu languages is normally treated as a morpheme/head located outside the argument structure domain (i.a.
(6)
Derivation of (5-a) (to be completed in (8))
Another relevant aspect of Ndebele clause structure is the fact that agreed-with objects undergo dislocation, to the right (7-a) or to the left (7-b) (
(7)
a.
UThabani
1Thabani
u-∅-
1s-
[vP
{*inyama} ]
9meat
kahle
well
{✓inyama}.
9meat
‘Thabani cooks the meat well.’
b.
Inyamai
9meat
uThabani
1Thabani
u-∅-
1s-
[vP
kahle.
well
‘The meat, Thabani cooks it well.
The derivation in (6) is thus not complete: it should involve a final step of object movement to the right periphery, shown in (8). I take this position to be a rightward specifier of a phrase above TP, whose head X triggers movement of DPs with the
(8)
Complete derivation of (5-a)
Since this paper is concerned with phenomena taking place inside TP, i.e. before object dislocation applies, I will represent objects as located in their base positions even when they control object agreement, keeping in mind that they undergo further movement once TP is merged with X. I return to the discussion of object dislocation in section 5.
Turning to the core data, subject agreement in Ndebele is typically
(9)
The assumed syntax of the argument structure domain
I use the term
(10)
U-Zodwa
a-1Zodwa
1s-
a-bantwana.
‘Zodwa made the children cook.’
(11)
U-Thabani
1s-
‘Thabani cooked.’
(12)
A-bantwana
2s-
‘The children were cooked for.
(13)
I-sihlahla
7s-
‘The tree fell.’
As previously observed (
(14)
*U-Zodwa
1s-
a-bantwana.
‘Zodwa made the children cook.’
(15)
*U-Thabani
1s-
‘Thabani cooked.’
(16)
*A-bantwana
2s-
‘The children were cooked for.’
(17)
*I-sihlahla
7s-
‘The tree fell.’
The canonical subject cannot control agreement on Voice even when it remains in situ (18)-(19).
(18)
a.
Kw-a-phek-a
15s-
u-Thabani
i-suphu.
‘Thabani cooked soup’
b.
*Kw-a-
15s-
u-Thabani
i-suphu.
‘Thabani cooked soup’
(19)
a.
Kw-a-w-a
15s-
i-sihlahla.
‘A tree fell.’
b.
*Kw-a-
15s-
i-sihlahla.
‘A tree fell.’
It is worth noting here that in-situ subjects in Ndebele are typically focused. Object agreement, on the other hand, targets non-focused DPs. Therefore, a possible explanation of (18-b) and (19-b) is in terms of discourse properties of in-situ subjects. This explanation would not, however, extend to the facts in (14)–(17), where the subject is not focused. Moreover, it would predict that agreement on Voice could be controlled by the
This pattern of agreement is a clear example of the NOM/ACC argument alignment: the highest argument in the clause is treated uniformly by the agreement system, irrespective of its thematic status. However, given the structure in (6), it is not clear how this pattern arises, i.e. why the probe in Voice ignores the closest DP in its c-command domain. Three types of analyses come to mind. The first would be to hypothesize that the object agreement probe is, in fact, not in Voice, but rather in a position from which it does not c-command the canonical subject. Second, this alignment might be due to an interaction between agreement and case that prevents Voice from agreeing with the subject. And third, the pattern could follow from the timing of subject movement and object agreement. In the next section, I entertain each possibility and discuss the challenges they face. Then, in section 4, I develop a version of the last type of analysis, arguing that the NOM/ACC agreement alignment in Ndebele is due to the interaction between subject movement and object agreement. I demonstrate that this interaction is, in fact, a consequence of an independently motivated parameter for Bantu languages, according to which ф-probes must cooccur with an
In this section, I present and reject three different ways of deriving the NOM/ACC agreement pattern described in the previous section: i) finding a position for the object-agreement probe that would derive of this pattern simply from probe-goal locality (section 3.1), ii) case-discrimination in agreement (section 3.2), and iii) countercyclic probing by T before Voice (section 3.3).
Recall from the previous section that the standard analysis of the object marker in Ndebele and closely related languages treats this morpheme as located outside of the argument structure domain (
(20)
Inflection [Macrostem OM [Stem Root + (extensions) + FV]]
Despite being well-motivated, the assumption that the object agreement probe is outside of the argument structure domain (here, in Voice) creates the puzzle laid out in the previous section: The object agreement probe in Voice is expected to agree with the canonical subject. In this section, I explore (and reject) the possibility that locating the probe lower in the structure might solve this problem and thus derive the NOM/ACC agreement alignment discussed above.
Locating the object agreement probe on little v, as has been proposed for many other languages,
Another incorrect prediction of this account is that Agents should always be unable to control object agreement. This is, however, not true. An Agent is unable to control object agreement
(21)
a.
Ngi-∅-
1
[vP tV
i-nyama ]
a-bantwanai.
‘I’m making the children cook meat.’
b.
Ngi-∅-
1
[vP
tV
a-bantwana
inyamaj.
‘I’m making children cook the meat.’
As argued by Zeller (
(22)
a.
Ngi-ya-
1
[vP
tV
kahle
well
a-bantwanai
i-nyamaj.
‘I’m making the children cook the meat well.’
b.
*Ngi-ya-
1
[vP tV ti
kahle
well
{i-nyamaj}
a-bantwanai
{i-nyamaj}.
‘I’m making the children cook the meat well.’
Next, observe that there can only be one object agreement morpheme on the verb (23).
(23)
*Ngi-ya-
1
[vP tV
kahle
well
a-bantwanai
i-nyamaj.
‘I’m making the children cook the meat well.’
Together, these facts show that there is only one object agreement probe in the Ndebele clause and that it can access the Agent of the verb
(24)
a.
A-bantwana
b-a-(
2s-
i-nyama.
‘Children cooked meat.’
b.
Kw-a-(
15s-
a-bantwana
i-nyama.
‘Children cooked meat.’
The same argument can be made from applied objects. An applied argument in an active clause (with a higher Agent) can control object agreement. But when the Agent is demoted under passivization, the same applied argument can no longer control agreement on Voice. What these facts show is not only that the probe is not in v, but more generally, that there is no single head to which we can fine-tune the position of this probe that would derive its inability to find the highest argument.
Let us turn to another possibility, which links agreement with case. It has been claimed for a number of languages that ф-agreement may be case-discriminating. For instance, Dative and Genitive DPs cannot control ф-agreement in many languages. A known example comes from Icelandic, in which Dative subjects do not control agreement on T (25).
(25)
Þeim
they.
var
be.
hjálpað
helped
‘They were helped.’ (Icelandic,
Similarly, Genitive subjects in Slavic languages like Russian and Polish cannot control subject agreement. The Polish example below shows the so called Genitive of Negation replacing Nominative marking on the subject in a negative clause (26).
(26)
a.
Oni
they.
byli
be.
w
at
domu. NOM → agreement
home
‘They were at home.’
b.
Ich
they.
nie
not
było
be.pst.
w
at
domu. GEN → no (default) agreement
home
‘They weren’t at home.’ (Polish)
What would it take for the agreement pattern in Ndebele to be explained in terms of case discrimination? The language would have to have a case system that assigns the same case to any DP that’s base-generated in the highest position in the clause. Let us call that case Nominative. The ф-probe in Voice (but crucially, not the one in T) would have to discriminate against this case (27).
(27)
[TP Tф [VoiceP Voiceф[-NOM] [vP DPNOM … ]]]
In the configuration above, ф in Voice probes first but, since it discriminates against Nominative DPs, it is not able to target the highest DP in the vP.
As straightforward as it may initially seem, there are two problems with this solution. First, known instances of case-discriminating agreement are ones where the probe is specified to look for a DP
A different possibility is to adopt the parametric setting that Bobaljik (
The second issue with the case-discrimination hypothesis is its deep incompatibility with what is known about case in Bantu languages, which have neither morphological case marking nor the kind of restrictions on DP distribution that have long been associated with abstract case in other languages (
In Halpert’s (
(28)
A-ku-phek-i
(u)-Zodwa.
(
‘Zodwa didn’t cook’
According to Halpert, Zulu has a two-case system. Structural case, licensed by a head between T and Voice, is exponed by the absence of an augment. DPs
(29)
a.
A-ngi-(*m)-bon-i
Zodwa
1Zodwa
‘I don’t see Zodwa’
b.
A-ngi-(m)-bon-i
‘I don’t see Zodwa’
However, the highest DP cannot control agreement on Voice irrespective of the presence of an augment (30).
(30)
a.
A-ku-(*m)-phek-i
Zodwa.
1Zodwa
‘Zodwa didn’t cook’
b.
A-ku-(*m)-phek-i
‘Zodwa didn’t cook.’
If Halpert’s theory is the correct theory of case in Ndebele (see
Analyzing Xhosa, another closely related language, Carstens & Mletshe (
(31)
Different cases borne by in-situ subjects in Xhosa (
a.
[vP AgentStructural Case ]
b.
[vP AgentFocus Case Theme ]
c.
[vP ThemeStructural Case ]
d.
[vP ExperiencerInherent Case ]
In such a case system, Voice would have to discriminate against all the cases in (31), including structural case. This incorrectly rules out agreement with objects of SVO clauses which, as Carstens & Mletshe propose, bear structural case as well. Moreover, there is evidence that Voice does not discriminate against the case borne by Experiencer arguments in Ndebele. First, Experiencer subjects, like all other subjects, can be either preverbal or postverbal. As with all other subjects, they cannot control agreement on Voice, whether moved to Spec,TP (32-a) or left in situ (32-b).
(32)
a.
U-Zodwa
u-(
1s-1o-sad-
‘Zodwa was sad.’
b.
Ku-(
15s-1o-sad-
u-Zodwa.
‘Zodwa was sad.’
However, the same Experiencer argument can control agreement on Voice if it is not the highest DP in the clause. This is the case when the verb ‘be sad’ (the same verb used in (32)) is causativized, which results in the presence of a higher argument, the Causer:
(33)
Ba-
2s-1o-sad-
u-Zodwa.
‘They made Zodwa sad.’
Thus, even if Experiencers bear a special inherent case, this case cannot be the reason why Voice cannot agree with it in (32).
Finally, many authors have independently argued that ф-agreement in Bantu languages is insensitive to case, irrespective of their view on the case system itself (i.a.
A third way to derive the NOM/ACC agreement alignment in Ndebele is by manipulating the timing of subject movement and object agreement. In particular, the subject would have to move out of the c-command domain of the object agreement probe before the probe initiates its operation.
It is typically assumed that the first step of A-movement that targets the subject is triggered by
(34)
Analyzing the same facts in Zulu, Zeller (
(35)
The first vP-external
As should be clear, this principle is a stipulation and should be avoided if possible, particularly if we want to maintain the restrictiveness brought about by assuming cyclicity.
As it turns out, it is possible to avoid (35). Note that the countercyclicity problem would not arise if the subject movement probe and the object agreement probe were both located in T. Such a derivation would be cyclic in the sense that no features of a higher head would probe before the features of a lower head. This type of analysis has been proposed to account for (apparent) countercylic effects in Icelandic experiencer constructions (
(36)
Ngi-za-{*yi}-be
1
ngi-sa-{yi✓}-phek-a.
1
‘I will still be cooking it.’
There is another way for the subject movement feature and the object agreement probe to appear on the same head: both could be in Voice. This, in turn, would mean that the first step of A-movement subjects undergo in this language is short, middlefield movement to Spec,VoiceP. In the remainder of this paper, I argue that this is indeed the case in Ndebele. In the next section, I lay out the details of the analysis and demonstrate how it derives the basic pattern of agreement in Ndebele. In section 6, I provide further evidence for this analysis by showing that it offers a straightforward explanation for three other puzzles: i) defective intervention effects in VSO clauses (6.1), ii) the optionality of movement to Spec,TP, (6.2), and iii) an object agreement asymmetry found in passives (6.3).
I propose that the NOM/ACC alignment in agreement in Ndebele is a consequence of parametric bundling of ф with
A well-known property of Bantu languages is the interdependence of agreement and movement. It is manifested robustly in the realm of canonical subjects, which must raise to Spec,TP in order to control agreement on T (37-a). In-situ subjects cannot control agreement (37-b).
(37)
a.
U-Thabani
1s-cook-
i-nyama.
‘Thabani cooks meat’
b.
*
1s-cook-
[vP
u-Thabani
i-nyama].
‘Thabani cooks meat’
When the subject stays in situ, T surfaces with class 15 agreement (38-a). Default agreement is not possible if the subject raises to Spec,TP (38-b).
(38)
a.
15s-cook-
[vP
u-Thabani
i-nyama].
‘Thabani cooks meat’
b.
*U-Thabani
15s-cook-
i-nyama.
‘Thabani cooks meat’
This one-to-one correspondence between movement and agreement is almost exceptionless in Bantu languages and it underlies proposals that link agreement with movement in this language family.
(39)
A ф-probe must cooccur with an
According to (39), whenever we see ф, there is also an
Recall that subject agreement in Ndebele is triggered by a ф-probe in T, and object agreement by a ф-probe in Voice. Due to the bundling parameter, both heads additionally bear
(40)
[TP T⟨ф,
Since arguments can stay in situ, the ⟨
(41)
Proposal
a.
T has ⟨ф:__,
b.
Voice has ⟨
c.
Voice is a phase head
With these details in place, we can derive the NOM/ACC agreement pattern in a way that does not involve countercyclicity, as detailed below.
An SVO sentence with object agreement, such as (42), is derived when Voice has ⟨
(42)
U-Zodwa
w-a-yi-phek-a
1s-
i-nyama.
‘Zodwa cooked the meat’
(43)
Step 1:
I assume that the
(44)
Step 2: The ф-probe in Voice, relativized to
Since
At this point in the derivation, object agreement has been established with the internal argument and the external argument has moved to the left edge of the VoiceP phase. In this position, the Agent is accessible to probes in T (before movement, the Agent was contained in the VoiceP phase):
(45)
Step 3: ф in T agrees with the Agent
Step 4:
The probe ordering in T gives rise to apparent spec-head agreement: T first agrees with the Agent in Spec,VoiceP and subsequently merges it as its own specifier.
SVO sentences without object agreement, such as (46), arise when ф in Voice does not find a matching goal, e.g. when the object does not bear an
(46)
U-Zodwa
w-a-pheka
1s-
i-nyama.
‘Zodwa cooked meat’ [Can answer: “What did Zodwa cook/do?”]
When ф in Voice fails the search of its c-command domain, it fails to agree altogether (see section 4.2 for discussion of why it does not agree with the subject in this case). The unvalued ф-probe in Voice has no overt exponence (47-a). This is different from an unvalued ф-probe in T, whose exponence is class 15 agreement prefix
(47)
a.
[Voice, ф] → ∅
b.
[ф] → /ɣu/(‘ku/kw’)
Now consider the VSO version of the same sentence (48).
(48)
Kw-a-(
15s-
u-Zodwa
i-nyama.
‘Zodwa cooked meat’
The in-situ subject is contained in the VoiceP phase, and so T cannot reach it. T’s
(49)
The structure of (48)
A similar analysis was proposed by Carstens & Mletshe (
(50)
(
a.
SVO: [TP T[+
b.
VSO: [TP T[-
On this account, an in-situ subject is only possible when T is defective. For Carstens & Mletshe (
(51)
defective T
While this account derives the fact that an in-situ subject cannot control object agreement (v does not have an agreement probe in VS orders), it does not explain why the subject cannot control object agreement in
(52)
Ngi-thembis-e
1
u-mama
[ uku-
i-ndlu
kusasa].
tomorrow
‘I promised mother to clean the house tomorrow.’
The temporal anchoring of infinitive in (52) is different than that of the matrix clause: the infinitive is future-oriented relative to the matrix clause. This strongly suggests that the infinitival clause is a TP (see e.g.
The present analysis derives the fact that A-movement and object agreement are not sensitive to the base-generation position of arguments. No matter where a DP originates, if it’s the highest, it will move out of the search domain of the object agreement probe before probing begins. This robust pattern in schematized in (53), where each circle corresponds to a different argument structure, each of which is illustrated in (54)–(57).
(53)
Possible controllers of ф-agreement with Voice in different argument structures
(54)
a.
y-a-(
9s-
u-Thabani.
‘The girl made Thabani cook.’
b.
Kw-a-(
15s-
u-Thabani.
‘The girl made Thabani cook.’
c.
I-nkazanai
y-a-
9s-
‘The girl made Thabani cook.’
(55)
a.
w-a-(
1s-
a-bantwana.
‘Thabani cooked for children.’
b.
Kw-a-(
15s-
a-bantwana.
‘Thabani cooked for the children.’
c.
U-Thabanii
w-a-
1s-
‘Thabani cooked for the children.’
(56)
a.
b-a-(
2s-
i-nyama
ngu-mama.
by-mother
‘The children were cooked meat by mother.’
b.
A-bantwanai
b-a-
2s-
ngu-mama.
by-mother
‘The children were cooked the meat by mother.’
(57)
a.
s-a-(
7s-
‘The tree fell.’
b.
Kw-a-(
15s-
‘The tree fell.’
This analysis derives the apparent asymmetry in the directionality of agreement triggered by Voice and by T. T always agrees with the DP in its specifier – a spec-head agreement pattern (58). In contrast, Voice always agrees downwards and cannot reach its specifier – a rigid downward Agree pattern (59). Both patterns are derived via the same operation, (downward) Agree, combined with variable timing relative to internal Merge.
(58)
a.
Description: head H always agrees with its specifier.
b.
Source: ⟨ф,
(59)
a.
Description: head H agrees either downward or not at all
b.
Source: ⟨
Note that the analysis presented so far only explains why the ⟨
A key feature of the proposed analysis is the ф→
In section 4.1, I proposed that
(60)
No goal for ф in Voice
Recall that, in the absence of a c-commanded goal, ф in Voice remains unvalued and is spelled out with a null exponent. One might expect, however, that, after failing to locate a goal in its c-command domain, the ф-probe would undergo cyclic expansion (
(61)
(62)
(63)
Note that if
I propose that the absence of probe expansion in the case at hand is ultimately due to the ordering of
(64)
Structural Description:
i.
nodes
ii.
probe feature F of
Structural Change:
As we will see immediately below, the crucial aspect of (64) is that probes expand only to root nodes. Let us see how this derives classic cases of cyclic Agree, such as those discussed by Béjar (
(65)
A cyclic Agree paradigm (Georgian, data from
a.
‘I draw
b.
‘
This cyclic Agree pattern follows from the definition of Probe Expansion in (64). When v is Merged with its complement, the
(66)
a.
Agree (failed)
b.
Probe Expansion
c.
Merge of EA
(66)-a meets the structural description for Probe Expansion: root node
(67)
a.
Merge of EA
b.
Agree (failed)
c.
No Probe Expansion
(67)-b does not meet the structural description for Probe Expansion: there is no
Such a bleeding derivation is responsible for the absence of Probe Expansion in VoiceP in Ndebele. As proposed above, Merge applies before Agree in Voice, creating a new root node and thus bleeding the expansion of the unvalued ф-probe. The only difference between Ndebele and the counterfactual Georgian scenario in (67) is that Merge in Ndebele VoiceP is internal, not external:
(68)
a.
Internal Merge
b.
Agree (failed)
c.
No Probe Expansion
Again, because only root nodes can inherit probes, Probe Expansion cannot percolate the ф probe from Voice to Voice′ in (68-c). Thus, if ф in Voice fails the search of its complement, it fails the search terminally – a rigid downward Agree pattern.
A noteworthy consequence of the proposed definition of Probe Expansion is that expanded probes can only agree with
(69)
Probe Expansion by context (Merge type and relative timing of Merge and Agree)
⟨Agree, Merge⟩
⟨Merge, Agree⟩
External Merge
Yes
No
Internal Merge
No
No
A further consequence of this view is that it allows us to treat expansion as a general property of probing, present even in languages/constructions which do not exhibit a cyclic Agree pattern. The lack of such a pattern could be due to the
As in related Bantu languages, object agreement in Ndebele requires object dislocation (see section 2.1). A common assumption is that, like subject agreement, object agreement is triggered by the head that attracts the agreeing DP to its specifier/adjunct position (e.g.
Before introducing the analysis, it should be made clear that dislocated objects in Ndebele are indeed moved, rather than base-generated in the dislocated position. Under the base-generation view, object agreement would be controlled by a
(70)
a.
U-∅-fun-ak
2
[vP
a.lot
[CP
ukuthi
COMP
*(A)-1who
a-buy-e?
1s-come-
]j
‘Who do you really want to come?’
b.
U-∅-fun-ak
2
[vP
1who
a.lot
[CP
ukuthi
a-buy-e?
1s-come-
]j
‘Who do you really want to come?’
The use of an interrogative pronoun as the raising DP helps us diagnose the landing position of raising-to-object. It has been previously shown that interrogative pronouns in Ndebele may lack the augment prefix (here
Returning to the main issue, I propose that right-dislocation is triggered by a probe located immediately above TP. This is motivated by the fact that right dislocation can target temporal adverbs, such as ‘yesterday’, which, by assumption, are base-generated in the tense-aspect domain:
(71)
[TP
Ngi-yi-phek-ile
1
{izolo}
yesterday
]
inyama
meat
{izolo}.
yesterday
‘I cooked the meat yesterday.’
Following Zeller (
(72)
Object right-dislocation in a transitive clause
We can now see why object agreement entails object dislocation, despite the fact that the two are not triggered by probes on the same head. The dislocation probe is relativized to
An immediate question arises: why can X access the in-situ object across the VoiceP phase boundary? This is made possible by the fact that the ф-probe in Voice is relativized to
(73)
Object dislocation:
Following Rackowski & Richards (
As we saw in section 3.1, multiple DPs may undergo dislocation in Ndebele. This suggests that X’s
(74)
U-Thabani
[VoiceP
u-ya-
1s-
a-bantwanai
i-nyamaj.
‘Thabani is cooking the meat for the children’
The dissociation of the object agreement probe and the dislocation probe makes it unsurprising that the direct object may be dislocated without controlling agreement, or more generally, that there can be multiple dislocated objects but only one instance of object agreement. ф-agreement is not what is responsible for object dislocation. ф in Voice is a simple (i.e. a non-insatiable) probe, which becomes inactive after being valued. The dislocation probe is an insatiable movement probe and does not require ф-agreement with its goals. It only requires access to the content of VoiceP, made possible by agreement with VoiceP itself.
Another correct prediction of this account is that subjects bearing
(75)
U-∅-phek-a
1s-
i-suphu
pandle
outside
u-Thabani.
‘Thabani cooks soup outside.’
As predicted, the insatiable probe on X will dislocate both the subject and the object if both bear
(76)
U-ya-yi-phek-a
1s-
pandle
outside
i-suphu
u-Thabani.
‘Thabani cooks the soup outside.’
Related to this is the generalization that dislocated subjects in Ndebele, as in many other Bantu languages, must control subject agreement:
(77)
*Ku-∅-pheka
15s-
i-suphu
pandle
outside
u-Thabani. (cf. (75))
‘Thabani cooks soup outside.’
The ungrammaticality of (77) falls under the generalization observed for many Bantu languages that vP-external subjects necessarily control subject agreement (see e.g.
Finally, we correctly predict that object dislocation is impossible if the subject remains in situ:
(78)
*Ku-∅-(yi)-pheka
15s-
u-Thabani
pandle
outside
i-suphu.
‘Thabani cooks soup outside.’
The ungrammaticality of (78) receives a similar explanation as (77). An in-situ subject entails that Voice lacks ф
(79)
Properties of right-dislocation and their explanation under the present account
i.
Explanation:
ii.
Explanation: The dislocation probe is relativized to
iii.
Explanation: Dislocation is relativized to
iv.
Explanation: The lack of subject agreement indicates that the subject is in-situ, which in turn means that VoiceP lacks ф
In this section, I discuss three puzzles in agreement and A-movement in Ndebele that receive a simple explanation under the proposed account of the agreement pattern discussed so far. These puzzles have been observed in other Bantu languages and they include: i) defective intervention effects in VSO clauses, ii) optionality of subject movement and agreement, and iii) an object agreement asymmetry in passives of ditransitives.
The account presented in section 4 derives the fact that in situ subjects do not control agreement on Voice, despite being c-commanded by it:
(80)
Kw-a-(
15s-
u-Thabani
i-nyama.
‘Thabani cooked meat’
In fact, Voice cannot agree with an object in this configuration, either:
(81)
Kw-a-(
15s-
u-Thabani
i-nyama.
‘Thabani cooked meat’
(80) and (81) together look like a case of defective intervention: the higher DP cannot control agreement on Voice and at the same time it blocks agreement with a lower, otherwise legitimate, goal. It is difficult to see, however, what would make the subject a
The analysis proposed here derives the ungrammaticality of (81) in a straightforward way. Recall that, in order for the subject to stay in situ, Voice must lack its
These facts additionally argue against an alternative analysis of why in-situ subjects do not control agreement on Voice: one might propose that an in-situ subject is not a possible goal for ф on Voice. Indeed, in-situ subjects are normally in focus and so they likely lack the
The present analysis provides a solution to another well-known property of many Bantu languages, namely the optionality of subject movement to Spec,TP and agreement with T. If T has
A different solution has been proposed for Zulu by Zeller (
(82)
a.
Kw-a-phek-a
15s-
u-Thabani.
‘Thabani cooked’
b.
*A-phek-a
u-Thabani.
‘Thabani cooked’
When T does not agree with the subject, its ф-probe is exponed as class 15 agreement. This, in turn, suggests that the probe is present but remains unvalued (
The other possible explanation of why sometimes T doesn’t interact with the subject was proposed by Zeller (
For Zeller (
(83)
a.
Kw-a-phek-a
15s-
[u-Thabani
kuphela].
only
‘Only Thabani cooked.’
b.
*[U-Thabani
kuphela]
only
w-a-phek-a.
1s-
‘Only Thabani cooked.’
On Zeller’s account,
(84)
a.
Ngi-fun-a
1sg-want-
ukuthi
[u-Thabani
kuphela]
only
a-phek-e.
1s-cook-
‘I want it to be the case that only Thabani cooks.’
b.
Ngi-fun-a
1sg-want-
ukuthi
ku-phek-e
15s-cook-
[u-Thabani
kuphela].
only
‘I want it to be the case that only Thabani cooks.’
(84) shows that movement to Spec,TP and agreement with T cannot be linked to discourse related features: the same kind of focus is available for the subject whether it moves to Spec,TP or stays in situ.
We have thus ruled out two alternative explanations of why T does not always find the canonical subject. This fact is not due to an optional absence of [ф,
The final, but important, piece of evidence for A-movement to Spec,VoiceP comes from object agreement in passives. Passives of double-object constructions are symmetric in that either object can move to Spec,TP and become the surface subject (85). Constructions in which the Theme (the lower of the two arguments) undergoes this movement will be referred to as
(85)
a.
y-a-phek-el-w-a
9s-
a-bantwana
‘The meat was cooked for children.’
b.
b-a-phek-el-w-a
2s-
i-nyama.
‘The children were cooked meat.’
One possible analysis of this kind of symmetry is A-scrambling of the Theme to a position in which it is a closer goal (or at least an equidistant goal) to T. Which head allows the scrambling? It is neither Appl nor v – each would overgenerate symmetric phenomena in the language. If the Theme could optionally scramble to the outermost specifier of ApplP, we would expect object agreement to be truly symmetric, contrary to fact (see section 3.1). If the Theme could scramble to the edge of vP, we would expect that the Theme could raise to Spec,TP across an Agent. While some Bantu languages, for instance Luguru (86), allow such Agent–Theme inversion constructions, this is not possible in Ndebele (87).
(86)
a.
Imw-ana
1-child
ka-tula
1s-broke
ici-ya.
7-pot
‘The child broke the pot.’
b.
Ici-ya
7-pot
ci-tula
7s-broke
imw-ana.
1-child
‘The child broke the pot.’ Luguru, (
(87)
#I-nyama
y-a-phek-a
9s-
u-Thabani.
#’The meat cooked Thabani.’
Cannot mean: ‘Thabani cooked meat.’
Thus, the scrambling of the Theme necessary for its passivization can be neither to the edge of ApplP nor to the edge of vP. Another possibility is scrambling to the edge of VoiceP. Under this hypothesis, inversion would arise due to the features of Voice. This offers a new way of understanding these facts: passive Voice can trigger A-scrambling, but active Voice cannot. This explains why inversion is possible in (85) but not in (87) – only the former is construed with Voice
Despite this symmetry observed in passives of ditransitives, there exists a puzzling asymmetry between them, described in (88) and illustrated in (89).
(88)
Object agreement asymmetry in passives
a.
In Benefactive passives, the Theme can control object agreement.
b.
In Theme passives, the Benefactive
(89)
a.
A-bantwana
b-a-
2s-
i-nyama.
‘The children were cooked meat.’
b.
I-nyama
y-a-(
9s-
a-bantwana.
‘The meat was cooked for children.’
The analysis proposed here accounts for both the symmetric behavior of the internal arguments as far as movement to Spec,TP is concerned, and their asymmetric behavior in controlling object agreement. In short, the symmetry in A-movement is due to the possibility of multiple specifiers of the passive Voice. The inability of the Benefactive to control object agreement in Theme passives is because the Benefactive is the highest argument in the clause – and those, under the present account, have no way of controlling agreement on Voice.
First, consider the derivation of a passive sentence with Voice having its usual ⟨
(90)
Benefactive passivization; Theme may control ф-agreement on Voice (derivation of (89–a))
When Voice
(91)
Theme passivization; Benefactive
Since internal Merge is ordered before ф-agreement in Voice, both movements take place before ф-probing: the first movement targets the Benefactive due to locality, the second targets the Theme. The reason why the Benefactive cannot control object agreement in Theme passives is because Theme passivization requires moving the Theme to Spec,VoiceP, which in turn entails that the Benefactive has moved to Spec,VoiceP, as well. DPs in Spec,VoiceP are invisible to the object agreement probe.
A conclusion that can be drawn for passives of ditransitives is that they are truly symmetric. The only asymmetry between them – object agreement – follows from a more general property of the language. It is an instance of the inability of the highest argument to control agreement on Voice, derived in a uniform way across different argument structures, including passives of ditransitives. Passives are especially interesting because Voice
The same object-agreement asymmetry in Zulu passives received a different treatment in Zeller (
A reviewer points out a potential shortcoming of this analysis: symmetric effects in passives and in object agreement are derived through different mechanisms and so the fact that they coexist in Ndebele is an accident. The two mechanisms are i) multiple specifiers of passive VoiceP for symmetric passives and ii)
A shared underlying mechanism is indeed more explanatory, but it is only warranted if the two phenomena – object agreement and movement to passive subject – are equally symmetric in a language. While this is true in many other languages, recall that the symmetry of object agreement in Ndebele is only an illusion. As discussed in section 3.1, the Benefactive blocks agreement with the Theme when both are matching goals (by virtue of having the
This paper addressed the question of why the thematically highest DP in Ndebele cannot control object agreement, despite being c-commanded by the object agreement probe. I argued that this exceptionless NOM/ACC agreement alignment can be understood as a consequence of the parameter that requires ф-probes to cooccur with an
I argued that alternative accounts in terms of case-discriminating agreement or based on probe-goal locality alone are untenable for Ndebele, and likely for many other Bantu languages, as well. This is not to say that case-discrimination and the position of the object agreement probe do not underlie superficially similar patterns in other languages. The position of probes is, naturally, predicted to be a factor in regulating argument alignment in agreement in all languages. We also independently know that case-discrimination in agreement exists and is (at least partially) responsible for the emergence of a NOM/ACC agreement alignment in some languages. However, we also know that the NOM/ACC agreement alignment is not restricted to languages with NOM/ACC case alignment (e.g. Warlpiri, Chukchi (
In fact, we expect the movement-based path to NOM/ACC agreement alignment to be empirically distinguishable from a case-based path. In a language in which the NOM/ACC agreement pattern follows from case, we may encounter instances of subject agreement with a DP that is not base-generated as the highest DP in the clause – for instance, when the highest DP bears a case that the subject agreement probe is unable to agree with. This is the case e.g. in Icelandic, where a Number probe may agree with a thematically lower DP if the thematically higher DP bears Dative case. This kind of agreement pattern is predicted to be absent in a language with a (purely) movement-based agreement alignment, like Ndebele. As we have seen, there is no property of DPs in Ndebele that would allow the thematically highest DP to “escape its fate” as the canonical subject in terms of agreement (that is, in its inability to control object agreement).
Finally, there appears to be no good reason to reject, as a matter of principle, the existence of multiple paths to the same argument alignment in agreement. With respect to case, it is generally agreed upon that what we call NOM/ACC alignment or ERG/ABS alignment are each a generalization that may emerge in different ways in different languages. Specifically, it is not the case that every language with an ERG/ABS case pattern has the exact same case system, if only because what we call Ergative is not an ontological primitive: in some languages, it is analyzed in configurational terms (i.a.
Unless otherwise noted, all data in the paper come from my own fieldwork.
Sentences with object agreement were elicited in a context where the object-marked DP is given in the prior discourse. For instance, the context for sentence (5-a) was the question “What’s going to happen to the meat?”
Zeller (
Pietraszko (
Given the obligatory dislocation of agreed-with objects, a possible analysis of object agreement in this language is as incorporated pronouns or clitics (
(i) Ngi-za-si-zam-a 1sg- uku-si-phek-a i-sitshwala. ‘I will try to cook porridge’
Finally, note that, even if object markers in Ndebele were clitics, the main question asked in this paper would still stand, although it would have a different form: Why is the highest argument invisible to the probe that triggers – not object agreement but – clitic movement. Under the hypothesis that clitics are licensed by ф-agreement (a common view in the literature on the PCC), this question could even be answered in the exact same way as I do in this paper: by the relative timing of subject movement and ф-agreement (that licenses the object clitic).
Exceptions to this are inversion constructions, including direct object passivization in double object constructions, discussed in section 6.3.
The precise structure proposed by Halpert (
(i) Ng-a-phek-is-a 1sg- a-bafana. ✓ a-2boys ‘I made boys cook.’
(ii) Ng-a-phek-el-a 1sg- u-mama. ✓ ‘I cooked for mother.’
(iii) *Ng-a-phek-is-el-a 1sg- {u-mama} { A-bafana {u-mama}. { Intended: ‘I made boys cook for mother.’
Given the absence of necessary evidence, I assume that the hierarchical order is uniformly Agent < Benefactive. Again, this detail is not of immediate relevance in this paper since no claims are supported by constructions in which the hierarchical swap of the Agent and Benefactive is observed in Zulu.
Note that the impossibility of object agreement with the subject in (14)-(19) is not due to a non-dislocated position of the subject. Preverbal subjects are ambiguous between being in an A-position and being left-dislocated. The post-verbal position is in turn ambiguous between being an in-situ or a right dislocated position. To ensure dislocation is not a confound here, I use distant past forms as they do not exhibit the so called conjoint/disjoint alternation. This alternation tracks argument dislocation in ways that can sometimes confound the evidence. Distant past verb forms are compatible with the following DP being in-situ or dislocated, thus eliminating the confound.
The status of the Stem as morphophological domain has been implemented in the syntactic literature has the result of head movement of the verb to the head exponed by the Final Vowel. Julien (
Some authors have assumed this for Bantu languages as well (e.g.
A reviewer points out that v in unaccusatives is often assumed to lack a ф-probe. This analysis accounts for the absence of object agreement in unaccusatives, but it does so by directly implementing the observation. The analysis I develop in this paper derives this observation as part of a broader phenomenon – the invisibility of the highest argument to the object agreement probe.
Condition C further shows that the indirect object c-commands the direct object:
(i) Ng-a-bon-is-a 1 yenaj him amaputha flaws ka-Thabanii/*j. of-Thabani ‘I made him see Thabani’s flaws.’ (Cannot mean: ‘I made Thabani see his own flaws.’)
A reviewer suggests another solution, namely that the object agreement probe is not always on the same head. Instead, Caus, v, and Appl can all be merged with a ф probe. This easily captures the facts: When the highest argument is a Causer, the probe would be in Caus; when the highest argument is an Agent, it’s on v; when the highest argument is a Benefactive, the probe is on Appl. As should be clear, this account is a restatement of the facts and so it does not provide an explanation for the NOM/ACC pattern that I seek here. Moreover, it incorrectly predicts the possibility of multiple object markers in a single clause since there is nothing that would a priori ban merging multiple heads with a ф-probe, at least in double-object constructions. For these reasons, I do not pursue this idea further.
Genitive of Negation in Polish typically replaces Accusative, with the exception of existential constructions such as the one in (26), where it replaces Nominative.
Even though the coocurrance of movement and agreement is very common in the Bantu family, not all Bantu languages can be characterized this way. See van der wal 2008 (
I do not commit to saying that all operations must be ordered in this way. A lack of specific ordering between
The different orderings in T and Voice can be understood as a difference between intermediate and final movement positions. Georgi (
I assume that an unchecked
A possible reanalysis of these facts would be that the infinitival T
(i) Ba-fun-a 2s-want- [uku-be be/*ku-hlabela.] 2s/*15s-sing. ‘They want to be singing’
If the subject of the infinitival clause were a PRO of class 15, we would expect class 15 agreement on the participle, contrary to fact.
A reviewer points out that this pattern may be characterized as antilocality: Voice cannot agree with the highest argument because the two are too close. The type of antilocality relevant here would be Erlewine’s (
This paradigm does not include a variant with an in-situ subject since speakers strongly prefer subjects of passive sentences to surface preverbally. I do not provide an explanation of this fact in this paper.
Béjar (
A similar claim has been made for Zulu by van der Spuy (
See section 4.1 for an argument against a yet different alternative, proposed by Carstens & Mletshe (
Strictly speaking, Zeller does not use probe relativization. Rather, he proposes that
The question of (a)symmetry in double-object constructions has been investigated extensively for Bantu languages. See, among others, Marantz (
Ndebele has locative inversion and instrumental inversion in active clauses, which may require a special kind of Voice to invert with subjects. I leave the analysis of those constructions for future work.
This analysis correctly predicts that in Theme passives, the Benefactive can be dislocated without object-marking, as evident below from the relative order of the Benefactive and a post-verbal adverb.
(i) I-nyama y-a-phek-el-w-a 9s- ti khale well a-bantwanai. ‘The meat was cooked well for children.’
A reviewer points out that
The Leipzig convention is used for glosses, with the following additions: 1 – class 1 nominal prefix (etc.), 1s – class 1 subject agreement (etc.), 1o – class 1 object agreement (etc.), A – augment vowel,
This research was conducted under a protocol approved by the Research Subjects Review Board at the University of Rochester.
For helpful comments, I’d like to thank Kenyon Branan, Mitcho Erlewine, Omer Preminger, and three Glossa reviewers. I’m especially grateful to my Ndebele consultants for providing data for this paper.
The authors have no competing interests to declare.