When grammatical constraints impose conflicting requirements on a linguistic expression, this conflict is often resolved by employing a repair operation. This repair can take various forms, for example insertion, deletion or modification of linguistic material. In this paper, I provide a number of case studies in morpho-syntax, showing how there are striking parallels between the repairs employed in phonology and morpho-syntax with regard to the context, type and shape of the repair. Ultimately, it is argued that, given clear similarities between these distinct domains of grammar, repairs should be governed by the same basic principles, namely a system of violable constraints. This view supports the hypothesis of Cross-modular Structural Parallelism (
The notion of “repair” is found across many domains of natural language, including syntax, morphology and phonology. Broadly speaking, a repair can be characterized as a particular (often generally unavailable) structural change licensed to avoid an illicit output configuration. Phonologists often treat processes such as epenthesis and deletion as repairs. For example, many languages do not allow for consonant clusters in the coda position of a syllable. If such a sequence arises, then this ill-formed structure is often “repaired” in some way. In Korean, one of the two consonants in a complex coda is deleted (1a). In Lebanese Arabic, on the other hand, an epenthetic vowel i is inserted to break up the offending cluster (1b). Each of these processes repairs the unwanted complex coda that would otherwise arise.
(1)
a.
/nəks/
/čəlm-ta/
→
→
[nək]
[čəm.ta]
*[nəks]
*[čəlm.ta]
‘soul’
‘young’
b.
/kib∫/
/ʔibn/
→
→
[ki.b
[ʔi.b
*[kib∫]
*[ʔibn]
‘ram’
‘son’
Morphosyntax is no stranger to repairs, either. A few representative examples of repairs in syntax involve
(2) | a. | do- |
[VP Read a book ] he |
||
I read a book and he did [VP — ] too. | ||
b. | ||
This is the man who1 I don’t believe [DP the claim [CP that anyone saw |
Repairs that are distinctly more morphological in nature are also frequently found. For example, there are numerous instances of haplology repair, involving dissimilation of sequences of adjacent homophonous morphemes (e.g.
(3)
a.
*Si
si
lava.
washes
‘One washes oneself.’
b.
si
lava.
washes
‘One washes oneself.’
Another morphological repair is the so-called
(4)
a.
Er
he
hat
has
das
that
gekonnt
can.
/
*können.
can.
‘He was able to do that.’
b.
Er
he
hat
has
das
the
Buch
book
lesen
read.
*gekonnt
can.
/
können.
can.
‘He was able to read the book.’
The question that this paper will address is whether it is possible to arrive at a general theory of repairs across domains. From a descriptive perspective, this seems to be a desirable goal, since repairs across domains share similar abstract properties, i.e. different repairs apply to the same marked output configuration, repairs have a “last resort” character and there is even intralinguistic variation with regard to the exact repair employed in a given context. In what follows, it will be argued that adopting violable constraints in phonology, morphology and syntax allows for a unified theory of repairs across domains that is able capture these cross-modular similarities. Furthermore, this lends support to the hypothesis of
(5) | |
Operations across distinct modules of grammar employ identical computational mechanisms. |
The central idea here is that modules of grammar should not differ in the abstract mechanisms they employ, but only in the alphabets that these operate on (e.g. morpho-syntactic vs. morpho-phonological features/structure). In addition, it will be shown that the assumption of violable constraints (as in Optimality Theory) allows us to be explicit about the following properties of repairs: (i) the context for repair, i.e. when a repair applies, (ii) type of repair, i.e. whether it involves addition, deletion or manipulation of a structure, (iii) the shape of the repair, i.e. what form the repair takes. As will be discussed, the few current conceptions of repair (such as “Last Resort”) remain undesirably vague and offer no principled explanations of any of the aforementioned properties. Theories with violable constraints, such as Optimality Theory, offer an explicit answer to all of these questions. While this has been previously been noted at several points in the literature (e.g.
Section 1.1 provides a brief introduction to violable constraints in Optimality Theory and Section 1.2 presents an example of the “Last Resort” conception of morpho-syntactic repairs and how this implicitly requires constraint violability. The following sections go on to illustrate how some of the core properties of repairs in phonology and morpho-syntax can be understood from an OT-perspective, namely their context (Section 2), type (Section 3) and shape (Section 4).
In Optimality Theory (OT) (
(6) | a. | |
C → Ø/___ ]σ | ||
b. | ||
Ø → V / C]σ ___ |
In OT, the effect of such rules is achieved by competition between potential output candidates, determined by the relative constraint profile of each candidate. There are two fundamental types of constraints in OT:
(7) | a. | N |
Syllables do not have codas. | ||
b. | D |
|
Do not insert. | ||
c. | M |
|
Do not delete. |
Competition between possible output candidates is represented in the form of a
(8) |
(9) |
In each of these cases, a “repair” such as deletion or insertion comes at the cost of violating a faithfulness constraint. However, the repair in question is licensed if this faithfulness constraint is deemed less important than the relevant markedness constraint (e.g. N
OT differs from most other linguistic theories in that competition is at the centre of all explanation. The determination of a well-formed expression is evaluated relative to other possible output forms, as prescribed by a set of ranked constraints. As such, OT is not a theory of phonology or syntax since it says nothing about the constraints themselves, only how they interact. This then opens the door to a general theory of repairs across domains – in each case, the challenge lies in identifying the relevant markedness and faithfulness constraints involved.
The spirit of violable constraints, although often not made explicit, can be identified in non-constraint-based approaches to repairs. By far the most widespread conception of a “repair” in morpho-syntax bears the moniker “Last Resort”. A definition of Last Resort is given in (10).
(10) | |
Operations must be driven by some condition on representations, as a “last resort” to overcome a failure to meet such a condition. |
In practice, however, the notion of Last Resort, when made explicit, is difficult to distinguish from a violable constraints approach to repairs. In fact, this was noticed early on by Prince & Smolensky (
In order to illustrate how the widely-adopted concept of Last Resort often tacitly involves OT-like concepts such as competition and constraint violability, let us consider the following Last Resort analysis from Bošković (
(11)
a.
On
he
je
is
ovladao
conquered
[NP
zemlj-om
country-
].
b.
*On
he
je
is
ovladao
conquered
[PP
with
[NP
zemlj-om
country-
]].
‘He conquered that country.’
Interestingly, if the complement of
(12)
a.
*On
he
je
is
ovladao
conquered
[QP
pet
five
zemalj-a
country-
].
b.
On
he
je
is
ovladao
conquered
[pp
with
[QP
pet
five
zemalj-a
country-
]].
‘He conquered five countries.’
Since (11b) shows that
(13) | a. | |
b. | ||
b’. |
This fits the profile of a repair in OT. Preposition insertion is not generally available, but can be used if the alternative would be even worse (i.e. unchecked case features). Bošković’s analysis can therefore be straightfowardly translated into an optimality-theoretic approach. Let us assume two basic constraints, the first is a markedness constraint F
(14) | a. | F |
Probe features ([ |
||
b. | D |
|
Do not insert prepositions. |
To capture the fact that preposition insertion is not freely available, it should outrank most markedness constraints, so that it usually lacks a trigger. However, in the case at hand, F
(15) |
However, in contexts where the NP bears genitive as the result of a particular numeral quantifier (12a), the option of agreeing with the NP in (16b) does not result in checking of the case probe on V and the fatal F
(16) |
Importantly, the violable constraints conception of this “Last Resort” repair is explicit about why this particular repair emerges and why its application is restricted to this context. In its simplest form, the logic is as follows: if a faithfulness constraint such as D
An important detail about Bošković’s analysis is that the prepositional phrase must bear an instrumental case feature [
(17)
a.
*Džokej
jockey
je
is
ovladao
conquered
Meri.
Meri
b.
Džokej
jockey
je
is
ovladao
conequered
with
Meri.
Meri
‘The jockey conquered Meri.’
c.
Džokej
jockey
je
is
pokušao
tried
ovladati
conquer.
our.
/
untamable.
Meri.
Meri
‘The jockey wanted to conquer our/untamable Meri.’
Translating the Last Resort analysis of Bošković into OT reveals some potentially problematic aspects of it. For example, there is presumably also a case probe for accusative on
In OT, repairs exist as competing derivational options that are almost always suboptimal in the unmarked case. However, in the few instances where this candidate is blocked by a higher constraint, lower-ranked constraints can have an effect in shaping the optimal grammatical output. In what follows, it will be shown that a general theory of repairs as the result of the fundamental OT tension between markedness and faithfulness allows us to have an explicit theory of repairs that encompasses the context, the type and the shape of repair operations in question. In particular, the role that violable markedness and faithfulness constraints play in driving and shaping repairs is summarized in (18).
(18) | ||
a. | The context for repairs is determined by output-oriented markedness constraints. | |
b. | The type of repair is determined by lower-ranked (faithfulness) constraints. | |
c. | The shape of repairs is determined by even lower-ranked markedness constraints. |
The following sections are devoted to a discussion of each of these aspects of repairs and how they can capture the striking similarities in repairs that we observe across the domains of phonology, morphology and syntax.
The first aspect of repairs to be discussed involves the context for repairs; in particular, the question of how the context for a repair operation is determined. We will see that the fact that repair operations converge on the same context lends support to the role of output-oriented markedness constraints.
An important argument for markedness constraints as the driving force for grammatical operations comes from what are known as
(19)
a.
/ʔilk-hin/
/lihm-hin/
→
→
[ʔi.l
[li.h
‘sing (aorist)’
‘run (aorist)’
b.
/lihm-al/
/ʔilk-al/
→
→
[lih.mal]
[ʔil.kal]
‘run (dubitative)’
‘sing (dubitative)’
Yawelmani also has a general process of word-final vowel deletion (20a). However, deletion is blocked if it would result in a complex coda, i.e. with affixation to a consonant-final base (20b).
(20)
a.
/taxaː-k?a/
/taxaː-mi/
→
→
[ta.xak]
[ta.xam]
‘bring!’
‘having brought’
b.
/xat-k?a/
/xat-mi/
→
→
[xat.k?a]
[xat.mi]
‘eat!’
‘having eaten’
Now, while it is possible to formulate two distinct phonological rules such as those in (21), this fails to capture the
(21) | a. | |
Ø → |
||
b. | ||
V → Ø /VC___# |
Epenthesis is
While rules such as those (21) fail to capture this, a reformulation in OT utilizing a markedness constraint such as *C
(22) | *C |
No complex syllable margins (*[σCC, *CC]σ) |
In accounting for epenthesis, we only require a corresponding faithfulness constraint D
(23) | |
The same set of the constraints also accounts for the fact that final deletion is blocked when it would result in a complex coda (20). In this analysis, final vowel deletion is driven by the markedness constraint *V# against word-final vowels. Since this constraint outranks the constraint against vowel epenthesis (M
(24) | |
What is important here is that the constraint *C
The first conspiracy involves the embedded CP domain in German. It is well-known that German is a V2 language requiring T-to-C movement if Spec-CP is overtly filled (
(25)
a.
Wen1
who
meinst
think
du
you
[CP
t1
[C′
[C0
dass
has
]
[TP
die
the
Maria
Mary
[
t1
getroffen
met
]
hat
]]] ?
b.
*Wen1
who
meinst
think
du
you
[CP
t1
[C′
[C0
Ø
]
[TP
die
the
Maria
Mary
[
t1
getroffen
met
]
hat
has
]]] ?
c.
Wen1
who
meinst
think
du
you
[CP
t1
[C′
[C0
hat2
has
]
[TP
die
the
Maria
Mary
[
t1
getroffen
met
]
t2
]]] ?
‘Who do you think (that) Maria met?’
Furthermore, German is known to have a construction in which extracted wh-phrases seem to be pronounced in multiple positions, sometimes referred to as
(26)
a.
Wer1
who
glaubst
believe
du
you
[CP
wer1
who
[C′
[C0
Ø
]
[TP
t1
[
Recht
right
]
hat
has
]]] ?
‘Who do you think is right?’
b.
Wer1
who
meint
said
Karl
Karl
[CP
wen1
who
[C′
[C0
Ø
]
[TP
wir
we
[
t1
gewählt
elected
]
haben
have
]] ?
‘Who does Karl say we have elected?’
While these two syntactic processes may seem unrelated, Fanselow & Mahajan (
(27) | *Ø |
Do not have a phonologically empty COMP domain (where both Spec-CP or C0 are empty). |
Thus, these operations are best viewed as repairs to the illicit representation in (28a). To avoid an empty COMP domain, either the C head must be filled lexically (28b), the verb moved to C (28c), or the copy of the wh-phrase in Spec-CP must be pronounced (28d).
(28)
a.
*Wen1
who
glaubst
believe
du
you
[CP
t1
[C0
Ø
]
[TP
Maria
Maria
[
t1
gesehen
seen
]
hat
has
]]
?
b.
Wen1
who
glaubst
believe
du
you
[CP
t1
[C0
dass
that
[TP
Maria
Maria
[
t1
gesehen
seen
]
hat
has
]]
?
c.
Wen1
who
glaubst
believe
du
you
[CP
t1
[C0
hat2
has
]
[TP
Maria
Maria
[
t1
gesehen
seen
]
t2
]]
?
d.
Wen1
who
glaubst
believe
du
you
[CP
wen1
who
[C0
Ø
]
[TP
Maria
Maria
[
t1
gesehen
seen
]
hat
has
]]
?
‘Who do you think Mary has seen?’
To phrase this in OT terms, we can postulate the following faithfulness constraints against head movement (29a) and copy deletion (29b), respectively.
(29) | a. | S |
Do not move heads. | ||
b. | C |
|
Lower copies in a movement chain are not realized. |
Given an input containing an empty COMP domain (as in embedded clauses from which extraction has taken place), the faithful candidate in (30a) fatally violates the high-ranked markedness constraint *Ø
(30) |
Interestingly, the option in (30d) of applying both T-to-C movement
(31)
*Wen1
who
glaubst
believe
du
you
[CP
wen1
who
[C0
hat2
has
]
[TP
Maria
Maria
[
t1
gesehen
seen
]
t2
]]
?
‘Who do you think Mary has seen?’
This supports the idea that these are actually repairs, since violable constraints require that repairs be as
(32) | T |
Function words are not pronounced (e.g. complementizers). |
However, some speakers of German, who do not have Doubly-Filled COMP effects in embedded clauses (
(33)
Wen1
who
denkst
think
du
you
[CP
wen1
who
[C0
(dass)
that
]
[TP
sie
she
t1
liebt
loves
]]
?
‘Who do you think she loves?’
This suggests that the lexical realization of the complementizer should not be treated as a syntactic repair. As for why (31) is possible and (33) is not, one possible account could be that lexical realization (i.e. Vocabulary Insertion) happens after optimization and the repairs. On this view,
Another example of a conspiracy in syntax involves the
(34)
a.
A
to
me
me.
importa
matter.3
solo
only
[pp
di
of
loro].
they.
All that matters to me is them.’
b.
A
to
me
me.
interessano
interest.3
solo
only
loro.
they.
‘I am only interested in them.’
The dative experiencer can also function as the antecedent for an object anaphor in each of these constructions. In (35a), the genitive object of the preposition now becomes the plural anaphor
(35)
a.
A
to
loro
them.
importa
matter.3
solo
only
[pp
di
of
se stessi].
themselves.
‘All that matters to them is themselves.’
b.
*A
to
loro
them.
interessano
interest.3
solo
only
se stessi
themselves.
‘All that matters to them is themselves.’
This led Rizzi (
As Woolford (
(36)
a.
Angutip
man.
arnaq
woman.
taku-vaa.
see-
‘The man sees the woman.’
b.
*Hansiup
Hansi.
immi
himself.
asap-puq.
wash-
‘Hansi washed himself.’
Inuit has two distinct ways of circumventing the AAE violation in (36b). The first is detransitivization of the verb, where the internal object is syntactically absent (37a).
(37)
a.
Asap-puq.
wash-
‘He washed himself.’
b.
Angut
man
immi-nut
himself-
taku-vuq.
see-
‘The man sees himself.’
Thus, we see that Inuit has two strategies that conspire to avoid agreement with an anaphor: either the anaphor is removed from the structure entirely, or it is assigned a case that makes agreement impossible. We can model this in OT as follows. Assume that there is a high-ranked markedness constraint against agreement with anaphors, i.e. the AAE (38a).
(38) | a. | A |
An anaphor may not control agreement. | ||
b. | A |
|
A head bearing a probe feature [F:□] agrees with a phase-local goal. | ||
c. | D |
|
Do not insert K heads. | ||
d. | M |
|
Do not delete DPs. |
Let us adopt a cyclic approach to optimization in which each step of the derivation is subject to optimization (see e.g.
(39) |
Since both M
The examples previously discussed serve to show that conspiracies exist both in phonology and syntax (also see
This section addresses the second major aspect of repairs in a theory of violable constraints, namely how one can account for which repair is chosen for a given context. It will be shown that the relative ranking of low-ranked faithfulness constraints that become active upon violation of a high-ranked markedness constraint determine the type of repair that a grammar opts for in a particular context. We will see this varies, as languages choose different repairs for the very same context. This will be illustrated by a comparison of hiatus repairs in phonology and PCC repairs in morpho-syntax.
When the repair for a particular configuration differs across languages, this is sometimes referred to as
(40) | ||||||
a. | Vowel elision: | /nd |
→ | [nd |
‘I build’ | |
(Xhosa) | ( |
|||||
/t |
→ | [t |
‘We have come’ | |||
(Chichewa) | ( |
|||||
b. | Glide formation: | /l |
→ | [l |
‘boats’ | |
(Luganda) | ( |
|||||
/a-r |
→ | [ar |
‘is eating’ | |||
(Okpe) | ( |
|||||
c. | Diphthongization: | /op |
→ | [o.p |
‘Lendu woman’ | |
(Ngiti) | ( |
|||||
d. | Glide insertion: | /dangl |
→ | [dangl |
‘to bevel’ | |
(Kalinga) | ( |
|||||
/m |
→ | [m |
‘middle- |
|||
(Faroese) | ( |
|||||
e. | ʔ-epenthesis: | /d |
→ | [d |
to lift ( |
|
(Malay) | ( |
|||||
f. | Coalescence: | /a-b |
→ | [ab |
‘seeds’ | |
(Foodo) | ( |
|||||
/mil |
→ | [mil |
‘it is red, they say’ | |||
(Tunica) | ( |
|||||
What is striking here is that the very same configuration can lead to a wide range of repairs. As well as providing further evidence for some universal markedness constraint determining the context of the repair (as with conspiracies), HoT/HoP effects such as this should inform the theory of repairs.
In general, this variation follows from two aspects of Optimality Theory: (i) violable faithfulness constraints, (ii) the assumption of a universal constraint set (CON). As shown above, repairs emerge with conflicts between markedness constraints (M) and faithfulness constraints (F), where M >> F. In this case, the violation of the lower-ranked faithfulness constraint (e.g. M
(41) | ||
Vowel elision | M |
|
Glide formation | *CG | |
Diphthongization | N |
|
Glide insertion/epenthesis | D |
|
Coalescence | U |
Assuming that the grammar of every language contains these in their constraint set, then the various repairs we find in (40) is determined by
(42) | |
*V.V >> N |
This is the most revealing case in which all relevant faithfulness constraints are ranked below the trigger anti-hiatus constraint *V.V. In this case, it is the candidate that violates the least costly (i.e. lowest-ranked) constraint, which will be selected as the optimal repair. Given the ranking in (42), this will be the insertion candidate (43d).
(43) |
The role of violable constraints is particularly important here. A repair always violates some constraint, however it is the lowest-ranked of these violations that ultimately determines the repair. Assuming universality of these constraints as well as re-ranking between languages allows us to capture both HoT/HoP effects and also conspiracies if these constraints are tied. The following section will show that similar arguments can be made on the basis of PCC effects in morpho-syntax.
There is a morpho-syntactic phenomenon with a strikingly similar HoT/HoP profile to hiatus contexts in phonology. This is what is known as the
(44) | ||
a. | ||
In a combination of IO |
||
b. | ||
In a combination of IO |
The PCC is designed to capture restrictions on the combination of certain “weak” elements (such as clitics, agreement affixes and pronouns) bearing certain person and case specifications. The original motivation for it comes from contrasts due to Perlmutter (
(45)
a.
Tha
stilune.
send.3
‘They will send him to you.’
b.
*Tha
stilune.
send.3
‘They will send you to him.’
As with hiatus, languages show a high degree of variability in the repairs they employ in PCC-violating contexts. A survey of some PCC repairs reported in the literature is given in (46).
(46) | |||||||||
a. | Substitution | IO3 | DO1/2 | → | LOC | DO1/2 | |||
(French) | ( |
||||||||
IO3 | DO1/2 | → | INANIM | DO1/2 | |||||
(Catalan) | ( |
||||||||
b. | Deletion: | IO1. |
DO2. |
→ | Ø | DO2. |
|||
(Ondarru Basque) | ( |
||||||||
IO2. |
DO1. |
→ | IO2. |
Ø | |||||
(Ondarru Basque) | ( |
||||||||
c. | Reflexivization: | IO3 | DO1/2 | → | IO3 | REFL1/2 | |||
(Georgian) | ( |
||||||||
d. | Case change: | DAT3 | ABS |
→ | DAT3 | ERG1/2 | |||
(Ondarru Basque) | ( |
||||||||
IO |
DO |
→ | IO |
DO |
|||||
(Finnish) | ( |
||||||||
e. | Preposition insertion: | IO3 | DO1/2 | → | [ |
DO1/2 | |||
(Kiowa) | ( |
||||||||
IO |
DO1/2 | → | [ |
DO1/2 | |||||
(French) | ( |
||||||||
IO |
DO1/2 | → | [ |
DO1/2 | |||||
(Catalan) | ( |
||||||||
f. | Metathesis: | IO3 | DO1/2 | → | DO1/2 | IO3 | |||
(Slovenian) | ( |
||||||||
Here, we see that languages opt for differing repair strategies for the same target context, and this is therefore a clear instance of HoT/HoP. Given the hypothesis of
Another interesting observation emerging from (46) is that, in PCC-violating contexts, the repair often affects the indirect rather than the direct object. This is a potentially surprising finding because, at least descriptively, the (strong) PCC imposes a more specific restriction on the DO and not the IO, i.e. that it cannot be a local person.
The final aspect of repairs to be discussed here involves what I will call the
If a language chooses deletion or modification as a repair, how is the target of this operation determined? To start with a morpho-syntactic example, numerous Romance languages do not allow for sequences of adjacent 3rd person clitics, sometimes called “3-3 effects” (
(47)
a.
*A
to
ella,
her
her.3
3
recomendé.
recommend.1
b.
A
to
ella,
her
3S
recomendé.
recommend.1
‘I recommended it to her.’
The traditional way of capturing this is by positing a special rule such as (48).
(48)
Delete/alter the features corresponding to 3rd person on a dativewhen it precedes another 3rd person.
However, such an approach does not tell us anything about why the indirect object is the target of this impoverishment rule, rather than the direct object. In her OT analysis of the
(49) |
On this view, the seemingly arbitrary choice of which element a repair targets in an OCP-like configuration is resolved by context-free markedness constraints. Since these markedness constraints are often ranked below faithfulness constraints, they generally do not have any influence output forms. However, in “default” contexts such as repairs, they exert their influence on the “shape” of a repair. Nevertheless, the target of repairs in 3-3 contexts could also be viewed as positional effect, that is, in a 3-3 sequence of clitics the rightmost one is protected from deletion or modification.
(50) | ||||
le | → | gli / | _____lo, la, li, le, ne | |
gli/le | → | ci / | _____mi, ti, vi | |
si | → | ci / | _____si | |
ne | → | ci / | _____ne |
As such, there is a potential ambiguity in the explanation of what determines the target of repairs to clitic sequences. Teasing apart these two options seems a worthwhile endeavour, but one that I leave to future research.
There are clear parallels to this in phonology. Recall from Section 3.1 that a possible repair in hiatus (V.V) contexts is deletion. However, there is still the question of which vowel is deleted. Casali (
(51) | ||||
/d |
→ | [d |
‘buy a cup’ | |
/ukp |
→ | [ukp |
‘yesterday’s cloth’ | |
/ow |
→ | [ow |
‘a different house’ | |
/umhel |
→ | [umhel |
‘some salt’ |
Thus, it is the position of the vowel that determines deletion (see
(52) | ||||
/t |
→ | [t |
‘I have them.’ | |
/m |
→ | [m |
‘He loves me.’ | |
/t |
→ | [t |
‘He howls it.’ | |
/t |
→ | [t |
‘I lead to him.’ |
Casali (
There are also cases of markedness-driven deletion in morpho-syntax. In the Ondarru dialect of Basque, the combination of clitics on the verb respects the PCC. In ditransitive configurations, the direct object can be third person (53a), but not first person (53b).
(53)
a.
Eur-ak
they-
su-ri
you.
Jon-
presenta
introduce
[d-
-
-
-Ø
-
-e]
-
(>tzue).
‘They introduced Jon to you (sg.)’
b.
*Eur-ak
they-
su-ri
you.
me-
presenta
introduce
[
-a
-
-
-
-Ø
-
-e].
-
‘They introduced me to you (sg.)’
In contexts such as (53b), some speakers repair this structure by omitting the absolutive clitic (54a). This results in the default linker morpheme
(54)
a.
Eur-ak
they-
su-ri
you.
me-
presenta
introduce
[
-
-Ø
-
-e]
-
(>tzue).
‘They introduced me to you (sg.)’
b.
Eur-ak
they-
ni-ri
me-
you-
presenta
introduce
[
aitu
-
_____
-Ø
-
-e]
-
(>satxue).
‘They introduced you to me (sg.)’
As Arregi & Nevins (
(55) | a. | |
b. |
This preference can be explained by positing two hierarchies of low-ranked markedness constraints for both person (56a) and case (56b), respectively.
(56) | a. | *1 >> *2 >> *3 |
b. | … >> *D |
Given the hierachies in (56), let us assume that these are ranked lower than the faithfulness constraint against deleted M
(57) | |
These case studies serve to show that low-ranked markedness constraints can be evoked to explain the preference in the target for a deletion repair.
As well as deletion, material can also be inserted as a repair. We previously saw examples in (1b) from phonology where epenthesis applies to break up cononsant cluster. To give another example, Kager (
(58)
/to-rm-n/
/apn-apn/
→
→
[tɔr.m
[ab.na.b
‘to his father’
‘free’
At this point, we face a similar question: How do we know which vowel to insert? In principle, there is a whole host of vowels that one could insert. As with deletion, this is determined by low-ranked, context-free markedness constraints.
(59) | |||
a. | [ə] | [–low, –round, +back, –high] | |
b. | [a] | [+low, –round, +back, –high] | |
c. | [ |
[–low, –round, +back, +high] | |
d. | [i] | [–low, –round, –back, +high] | |
e. | [u] | [–low, +round, +back, +high] |
Broadly following Kager (
(60) | |
In general, such constraints will be ranked below faithfulness constraints, so as not to have an effect on outcome.
Again, we find cross-modular parallels of this in the domain of morpho-syntax. A pertinent example involves the phenomenon of “default agreement”.
(61)
Marija
Marija
je
be.3
kupil-a
buy-
knjigu.
book
‘Marija bought a book.’
While this agreement tracks the
(62)
a.
Hladn-o
cold-
je.
is.3
‘(It) is cold.’
b.
Trebal-o
needed-
je
is.3
da…
that
‘(It) was necessary that…’
c.
Činil-o
seemed-
mi
me.
se
da…
that
‘(It) seemed to me that…’
We can interpret this as insertion of feature values not present in the structure, similar to epenthesis. The analysis requires constraints that we are already familiar with: A
(63) | a. | A |
T agrees with a locally-available goal. | ||
b. | D |
|
Do not insert feature values not present in the input. |
As with epenthesis, all insertion candidates will violate D
(64) | a. | *1 >> *2 >> *3 |
b. | * |
|
c. | * |
In the analysis, the values inserted are those contributing the least marked values given (64). This is the 3rd singular neuter form in (65b).
(65) | |
Since the choice of what material an insertion repair uses seems to relate to “unmarked” material, and can show some variance, appealing to low-ranked markedness hierarchies allows us to capture the parallelism between phonology and morpho-syntax.
A final example of how markedness constraints can shape repairs comes from what I will refer to as
(66)
Base
-
/aral/
um-aral
‘to teach’
/akyat/
um-akyat
‘to climb’
/bagsak/
b-um-agsak
‘to fail’
/sulat/
s-um-ulat
‘to write’
/gradwet/
gr-um-adwet
‘to graduate’
/preno/
pr-um-eno
‘to brake’
The classic analysis of the phenomenon treats the infix -
(67) | a. | A |
The affix - |
||
(Assign a violation mark for each segment between - |
||
b. | N |
|
Syllables do not have codas. | ||
(Assign a violation mark for each syllable with a coda) |
It is the conflicting requirements of these constraints that lead to repairs that find the best compromise for these two constraints. To see this, consider what happens when -
(68) | Vowel-initial base |
Things are more interesting if the base is consonant-initial. In (69a), placing -
(69) | |
The intuition of this analysis is therefore that we try to place -
The question now is whether we can find parallel examples in syntax. I will argue that multiple wh-movement in Romanian presents us with such a case. Consider first that Romanian is a multiple wh-fronting languages in which all wh-phrases are moved to the left-periphery (70).
(70)
[
Cine1
who
cui3
to.whom
ce2
what
[
ziceai
say.2
[
că
that
t1
i-a
to-him
promis
promised
t2 t3 ]] ?
‘Who did you say promised what to whom?’
Furthermore, if a wh-phrase is embedded inside a Complex NP Island, movement to Spec-CP of the matrix clause is blocked:
(71)
*[CP
Cine1
who
what
[
t1
o
cunoaşte
know
[
pe
studenta
student
[
căreia
which.
i
s-a
dedicat
dedicated
t2
ieri
yesterday
]]]] ?
‘Who knows the student to whom
What is more, Cheng & Demirdache (
(72)
*[CP
Cine1
who
[
t1
o
cunoaşte
know
[
pe
studenta
student
[
căreia
which.
i
s-a
dedicat
dedicated
what
ieri
yesterday
]]]] ?
‘Who knows the student to whom
A surprising fact, however, is that moving the wh-object
(73)
[
Cine1
who
[
t1
o
cunoaşte
know
[
pe
studenta
student
[
căreia
which.
what
i
s-a
dedicat
dedicated
t2
ieri
yesterday
]]]] ?
‘Who knows the student to whom
The puzzling question at this point is why moving to the edge of an island is grammatical, but remaining
(74) | a. | W |
[wh]-marked items must be in the specifier of a licensing head (C[wh]). | ||
(Assign a violation unused landing site between a wh-phrase and (including) its final landing site in Spec-C[wh]) | ||
b. | I |
|
Movement out of an island is prohibited. |
In the analysis, we see a strikingly parallel to the Tagalog example (75). Moving to Spec-CP as in (75a) fully satisfies W
(75) | |
This derives what Kotek (
This paper has shown that the way in which conflicts are resolved within a grammar, that is with various kinds of repairs, exhibits a striking convergence across domains. It was argued that the assumption of violable constraints (both in the form of markedness and faithfulness constraints) allows us to account for cross-modular parallels in the context, type and shape of repair operations. It was shown that linking the context of repairs to high-ranked markedness constraints can account for why, even when repairs may vary, they often converge on the same banned output context both within and across languages, as was demonstrated on the basis of conspiracies and HoT/HoP in both syntax and phonology. Furthermore, the OT conception of repairs results from a fundamental tension between markedness and faithfulness constraints. If some relevant faithfulness constraint is ranked below the relevant markedness constraint, then a candidate violating that constraint (e.g. deletion) emerges as a potential repair. If there are numerous constraints ranked below the relevant markedness constraint, it is the lowest-ranked of these that ultimately determines the repair. Variation in this regard can lead to HoT/HoP effects. Finally, the exact choice of what to delete, insert or modify is guided by even lower-ranked constraints. These effects tend to lead to preservation or insertion of the least-marked material possible.
This overall picture can, in a somewhat idealized form, be depicted as in (76). A markedness constraint M1 determines the context of the repair, the lowest-ranked faithfulness constraint F3 determines the type of the repair and the shape of the repair comes from the lowest-ranked markedness constraint M4.
(76) | |
As we have already seen, things are often more complicated than this. Markedness constraints can also help to determine the type of a repairs (e.g. N
A reviewer raises the question of whether it is not equally possible to account for the various aspects of repairs discussed here in a rule-based theory. For example, given the basic rule schema in (77), we could talk about the target (A), change (B) and context/environment (C) in which repairs take place.
(77) |
While I do not have sufficient space to devote a detailed discussion to this point, it seems apparent that this approach would seem to lack explanatory power with regard to the shape of repair. As we saw, this is very often driven by markedness considerations, whose place is less clear in a rule-based theory without markedness/faithfulness constraints. Furthermore, rule-based approaches still suffer from the perennial problem of capturing the
This paper has argued that an optimality-theoretic approach to repairs allows for a more explicit theory of repairs encompassing the context, target and shape of repairs. While the OT view of repairs ultimately does not tell us
1 = first person, 2 = second person, 3 = third person,
Rezac (
A remaining challenge for this analysis is to explain why
(i)
[VP V[
] [PP P[] [QP Q[*]
Consequently, it seems that the fact that
A reviewer mentions another domain in which a similar analysis is possible, namely nominalizations. An ordinary verbal predicate such as
(i)
a.
They destroyed (*of) the city.
b.
Their destruction *(of) the city.
c.
God declared (*of) them to be wrong.
d.
God’s declaration *(of) them to be wrong.
While this
I am assuming that tied optima results in optionality between the relevant outputs. As an anonymous reviewer correctly remarks, this is not necessarily an innocuous assumption. The treatment of ties in OT is discussed at length in Müller (
I would like to thank Sandhya Sundaresan for pointing out the following data as an example of a conspiracy.
It is important to mention that this is not an instance of anti-passivization. This is a distinct construction in which an antipassive morpheme -
(i) a. Tuqut- kill- ‘He killed.’ b. Tuqqur-puq. hide- ‘He hid himself.’
Two anonymous reviewers point out that postulating AAE as a violable constraint would predict that there should be languages with the ranking A
(i) a. Raaj-e Raj. potaa-ne self- sandov- involved- ‘Raj involved himself.’ b. Sudhaa-e Sudhaa. potaa-ne self- sando- involved- ‘Sudha involved herself.’
This does not seem to exhaust the possible repairs to the AAE cross-linguistically, there is also “agreement switch”, in which an object probe targets the subject just in case object agreement would violate the AAE (i.e. in Kutchi-Gujarati;
Of course, a single language can also have multiple ways of resolving hiatus (i.e. a conspiracy), see Baković (
It is important to note that there other kinds of PCC have been proposed since Bonet (
Note that this range of variation seems difficult to capture in the repair system suggested by Rezac (
In Cyclic Agree theories such as Béjar & Rezac (
Thanks to Andrew Nevins for making me aware of this point.
There is a technical difference, however, where Casali (
Kager (i/ is chosen for insertion (i).
(i) /t-n-ak-ol/ /ark-ark/ → → [t [ar.ga.r ‘you will do it’ ‘to growl’
ii
Kager (
For the sake of exposition, only the markedness violations pertaining to the epenthetic segment have been included.
This is why not all inputs are neutralized to the most unmarked form, e.g. [ba], as is sometimes asserted (e.g.
Another empirical domain that can be analyzed in this way is “default case” (e.g.
It should be noted, however, that the use of gradient alignment constraints has proven controversial, including for the Tagalog case presented here (
This is, of course, requires further refinement and potential decomposition into further, more specific constraints. Note that I assume that Ross-type islands such as the CNPC are representational constraints that hold at PF. Evidence for this comes from the fact that violations of representational islands seem to be repaired by operations that alter the offending representation, e.g. ellipsis (e.g.
There is still the question of whether one can find arguments for Cross-modular Parallelism from semantics/pragmatics. It certainly seems to be the case that aspects of semantics/pragmatics are optimality-theoretic in nature, as has been pointed out (e.g.
Of course, the fact that OT is possibly the only theory that makes explicit predictions about the extent of typological variation leads to an overgeneration issue sometimes referred to as the
This paper has benefited immensely from comments and suggestions from the editors of this Special Collection, Jana Willer-Gold and Andrew Nevins, as well as three anonymous reviewers. I would also like to thank the audience at Resolving Conflicts Across Borders 2017 and SinFonIJA X in Dubrovnik, Croatia. Further thanks go to Gereon Müller, Fabian Heck, Bob Frank, Sandhya Sundaresan, Gurujegan Murugesan and Joanna Zaleska for discussions of various aspects of this work at different stages of its development.
The author has no competing interests to declare.