Voice in Northern Pomo1 Edwin Ko April 23rd , 2015 Abstract 1

Voice in Northern Pomo 1
Edwin Ko
April 23rd, 2015
c. madal
(*pʰow) k′a:
3fs.Acc
3pl.Nom leave.PST
“She was left.”
d. mo:wal
pʰidim -ka
3ms.Acc hang
-caus
“(They) are hanging him.”
Abstract
The Northern Pomo ‘voice’ marker -ya (also -ʔa post-nasal) when applied to
intransitive verbs demotes the external argument, forming impersonal sentences.
This marker can also be applied to transitive verbs, and its external argument must
be unexpressed. In my paper, I will first introduce this voice marker in Northern
Pomo before examining previous works on the treatment of passives. This work is
largely inspired by Harley (2013)’s argument for the distinction between VoiceP and
vP using data from Hiaki. In contrast to Hiaki, the voice marker in Northern Pomo
does not promote the internal argument, instead retaining accusative case (if the
internal argument is expressed). Therefore, the question is whether constructions
containing -ya are cases of (a) passive voice without promotion, or (b) active voice
with null impersonal subjects. Using Maling (1993)’s diagnostics, I will demonstrate
that these tests provide inconclusive and ambiguous results, and provide an analysis
on binding of anaphors to suggest that these impersonals are actually active.
1 Background: Northern Pomo
Northern Pomo is an indigenous language of California of the Hokan stock.
1.1 Overview of Northern Pomo Syntax (O’Connor 1992)
-
Verb-final, fairly free order of arguments, and arguments may be unexpressed.
Heads follow dependents, or exhibits left-branching tree structure.
Verbs do not appear to agree for number or person. 2
Verbs have three (productive) valence-changing suffixes: causative, reflexive,
voice. In general, the order of these suffixes is causative » reflexive » voice.
-ya
-ya
According to Kiparsky (2013)’s OT typology, Northern Pomo is among languages
such as Finnish that (a) have intransitives passives, (b) disallow promotion, and (c)
disallow by-phrases.
2 Selected History of Passives
2.1 Burzio’s Generalization (1986)
Burzio’s Generalization is the observation that:
(i) A verb can only assign accusative case when an external argument is realized.
(Burzio 1986: 178-9)
(ii) A verb can only theta-mark an external argument when accusative case is
assigned.
(Burzio 1986: 184)
This is represented schematically below:
𝐴𝐴 ↔ 𝜃𝜃
𝑤𝑤ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝐴𝐴 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶, ∅ = 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑇𝑇ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅
(Burzio 1986: 184)
Problem: Northern Pomo assigns accusative case and lacks an external argument.
2.2 Case Absorption (Jaeggli 1986; Roberts 1987)
-
1.2 Northern Pomo -ya constructions 3
-
(1) a. loʔka
-ya
drop.PST -ya
“(It) was dropped.”
(2)
b. (*pʰow)
mu:
k′otam
3pl.Nom there
swim
“People swim there.”
-ya
-ya
The passive verbal suffix -en receives case and theta-role from the verb. In
other words, -en absorbs case and theta-role assignment. 4
Absorption is motivated by government; the verb governs the passive
morpheme. Below is the representation of the passive particle:
-ʔa
-ya
(Jaeggli 1987: 592)
1
Special thanks goes to Cathy O’Connor for allowing me to use her Northern Pomo data, as
well as sharing her wealth of knowledge on the complexity of the language.
2
Some verb stems undergo suppletion when the subject is plural (or collective), but this does
not entail that there is agreement for number.
3
The implicit argument must be human, and bear a [+human] feature.
Jaeggli (1986) claims also that passives containing by-phrases (i.e. long passives), the passive
suffix transmits the theta-role to the PP by-phrase. The theta-role percolates to the element by
which then assigns the external theta-role to the DP.
4
1
Problem: Northern Pomo verbs may contain several suffixes. In cases like (1d) and
assuming the syntactic structure in (2), the verb would not govern the voice marker
due to the presence of the causative suffix.
(4)
2.3 VoiceP (Kratzer 1996)
(Collins 2005: 96)
Voice is a functional category where external arguments are generated.
Agent comes from Voice, and Theme comes from the verb. 5
This configuration is represented below:
In order for the direct object of a transitive verb to move over the external argument
(i.e. the argument of the by-phrase), the concept of smuggling is used, and preferred
over head movement – the preposition by would have been in the way.
(3)
(Kratzer 1996: 132)
“Suppose there are two kinds of voice heads in English: active and non-active.
Active voice heads add external arguments and assign (check) accusative Case.
Non-active voice heads do not add external arguments and do not assign (check)
accusative Case.” (Kratzer 1996:123, emphasis added)
Problem: Again, Northern Pomo assigns accusative case and lacks an external
argument. However, the notion of VoiceP shows great potential.
(Collins 2005: 97)
The syntactic structure for the book was written is represented below. Notice that the
empty category in [Spec,vP] has null case that is checked by the null Voice°.
(5)
2.4 Smuggling (Collins 2005)
-
Proposes that actives and passives have the same syntactic structure as they
have the same argument structure.
VoiceP does not host the Agent, instead Agent is in [Spec,vP]. VoiceP only
present for passives. In other words, VoiceP = PassiveP.
In by-phrases, the preposition by is in the Voice° head.
“I propose that the external argument is merged into the structure in the passive in
the same way as in the active. In particular, I propose that the external argument is
merged into Spec,vP in the passive, just as the external argument is merged into
Spec,vP in the active” (Collins 2005: 84)
(Collins 2005: 102)
Problem: Predicate fronting, or smuggling, is not necessary in Northern Pomo
because (a) the language is verb-final, and (b) the internal argument is not promoted.
Furthermore, the presence of an external argument as either Ø or pro is ambivalent
5
Kratzer (1996) distinguishes between two kinds of active Voice° - eventive and stative. Here,
I subsume both varieties under the same category.
2
in -ya sentences which will be discussed later – the question is then whether it is
even necessary to encode the external argument as an empty category in [Spec, vP].
The motivation for structural symmetry between use/non-use of -ya is very much
unclear.
2.5 vP ≠ VoiceP (Harley 2013, 2014a)
-
-
Attempts to address (a) the base-position of external arguments, and (b) how
these arguments are ‘demoted’ or ‘suppressed’ by morphology like the passive.
Argues for a distinction between VoiceP and vP (and ApplP complement to vP,
if present) – VoiceP is higher than applicatives/causatives according to the
order of morphemes (cf. Mirror Principle).
Proposes VoiceP in addition to vP and VP where Voice° head, with a vP
complement, introduces the external argument (instead of v°).
-
The external argument of hi’ibwa is base-generated in [Spec, VoiceP] before
promotion, or movement to [Spec, TP].
The causative and verbalizing morphemes are in the v° heads.
Implicit argument not included into the syntactic representation.
Problem: Unlike Hiaki, the voice marker in Northern Pomo does not promote the
internal argument. Furthermore, Harley excludes external arguments as Ø/pro from
the syntactic diagrams, and concludes that “this variable is bound via what
Wurmbrand (2013) calls ‘semantic control’” (Harley 2013). As mentioned, this may
be a problem for Northern Pomo in establishing whether the voice marker signifies
impersonal passive or impersonal active.
3 VoiceP, vP and VP in Northern Pomo
Mirror Principle: The order of morphological operations mirrors the order of
syntactic operations.
(Baker 1985)
I argue that Northern Pomo provides further evidence supporting verb phrases being
tripartite, made up of VoiceP, vP and VP/√P. Northern Pomo and Hiaki share very
similar morpheme ordering seen in (7) below:
Monotonicity Hypothesis: morphology can only add to the syntactic structure, not
delete or remove from it.
(Koontz-Garboden 2007)
(7)
mo:wal
pʰidim -ka
3ms.Acc
hang
-caus
“(They) are hanging him.”
-ya
-ya
Passive marker in Hiaki:
No external arguments may appear, nor can they appear in a by-phrase.
Able to apply to both intransitive and transitive verbs, but requires promotion
of internal argument (if present) to subject position.
Implicit agent always [+human].
Passive of a causative can be seen below:
(6)
3.1 Unaccusative verb + -ya
(8)
loʔka
-ya
drop.PST -ya
“(It) was dropped.”
6, 7
(Harley 2013: 53)
6
√P is used here synonymously with VP. Some scholars use √P, especially in Distributed
Morphology, to identify the root, “pure units of structural computation” (Harley 2014b).
The head v° containing Ø represents verbalizing morphology in Hiaki (i.e. cho’oko “salt” →
cho’okote “to salt”).
7
3
3.2 Unergative verb + -ya
(9)
(*pʰow)
mu:
k′otam
3pl.Nom there
swim
“People swim there.”
4 Diagnosing -ya
-ʔa
-ya
Maling (1993) and Maling & Sigurjónsdóttir (2002): “‘impersonal passives’ of
intransitive verbs are in principle syntactically ambiguous between an active or a
passive analysis.” (Maling 2015) Maling considers the crucial difference to be the
occupant in the subject position of impersonals:
Passive: In subject position, an empty category [e].
Active: In subject position, “a syntactically effective thematic subject” [proarb].
Maling (1993; 2015)’s diagnostics can be represented in the following table:
Syntactic Property
Agentive by-phrase
possible
Binding of anaphors by
null ext. arg. possible
Control of subjectoriented adjuncts by null
ext. arg. possible
Non-agentive
“unaccusative” verbs
possible
3.3 Transitive verb + -ya
(10)
madal
(*pʰow) k′a:
3fs.Acc
3pl.Nom leave.PS
“She was left.”
-ya
-ya
-
How does the internal argument get accusative Case?
Perhaps the Voice° containing the suffix -ya preserves its Case feature, in contrast to
Hiaki which lacks it.
What about EPP?
Using the OT analysis by Kiparsky (2013), expanded from Ackema and Neeleman
(1998), the EPP constraint – a sentence must have a thematic subject – is dominated
by PARSE (i.e. the input is realized by the voice marker), and STAY (i.e. the subject
bears the most prominent theta-role). Another line of thought is to simply say that
either the TP does not bear an [EPP] feature when voice marker is present.
Should the external argument be excluded from the syntactic structure?
It is still not clear whether the voice marker in Northern Pomo is an active
impersonal or a passive impersonal. The demarcation of active vs. passive
impersonals has been a topic of wide debate among linguists which has been
addressed by Maling (1993; 2015). This is discussed in the next section.
-
Active Clause
Passive Clause
✓
✗
✓
✗
✓
✗
✗
✓
The “autonomous” construction in Irish (McCloskey 2007) and the -no/-to
construction in Polish (Maling & Sigurjónsdóttir 2002) vs. Ukranian (Sobin 1985)
exhibit similar impersonal constructions.
Maling (2015) applies her diagnostics to these constructions:
Syntactic Property
Agentive by-phrase
possible
Binding of anaphors
by null ext. arg.
possible
Control of subjectoriented adjuncts by
null ext. arg.
possible
Non-agentive
“unaccusative” verbs
possible
Polish (=active)
Ukranian (=passive)
Irish (=active)
✓
✗
✓
✓
✗
✓
✓
✗
✓
✗
✓
✗
4
O’Connor & Maling (2014) applied these diagnostics to Northern Pomo 8:
4.4 Non-agentive “unaccusative” verbs
4. 1Agentive by-phrase
-
-
(17)
As mentioned, agentive by-phrase not possible.
4.2 Binding of anaphors
-
Emphatic reflexive k’aye cannot be controlled by the subject.
(11)
-
*k’aye
ma:dal mako
-ya
self
her
find
-ya
* “They THEMSELVES found her.”
loʔka
-ya
drop.PST -ya
“(It) was dropped.”
We therefore get the following ambiguous results from these tests:
Syntactic Property
Agentive by-phrase
possible
Binding of anaphors by
null ext. arg. Possible
Control of subjectoriented adjuncts by null
ext. arg. Possible
Non-agentive
“unaccusative” verbs
possible
* “She HERSELF was found.”
Subject-controlled non-clause bound ‘logophoric’ reflexive pronouns (NCBR)
possible in long-distance.
(12)
Non-agentive “unaccusative” verbs possible.
tiyaʔ
diyi
ma:dal
čima -ka
-ya
NCBR.Obl.pl next
her
sit
-caus -ya
“She was made to sit next to them.”
…but not in short-distance…
Northern Pomo
✗
?
?
✓
(13) *tiyaʔ
hayunam čaban -ʔa
NCBR.Obl.pl dog.spec kill
-ya
* “They killed their own dog.”
If actually active, then…
4.3 Control of subject-oriented adjuncts, and of embedded subjects
1. Perhaps the [EPP] feature is present, satisfied by pro.
2. No need to assume case preservation – like Hiaki, Voice° lacks Case feature.
-
Control of adjunct subjects possible.
(14)
pʰiʔelaʔ -hI
(*pʰow) ma:dal mako -ya
seek.pl -Acomp they her
find
-ya
“Having searched, they found her.”
…but control of subject w/ control verbs not possible...
(15)
*ma:dal [ duhu -ka ] daʔan -ʔa
her
leave -caus want -ya
* “She was wanted to leave.”
…nor can passive clauses occur as complement to control verb…
(16)
*mo:w [ ma:dal čaban -ya ] daʔad -e
he
her
kill
-ya
want -pres
* “He wants her to be killed.”
5 Reanalysis of bound anaphora
Northern Pomo does not have a fixed case-marking system, instead there is evidence
that Northern Pomo is a Fluid S Case-Marking system (Deal & O’Connor 2010).
(18) k’aye ma:dal šičuka
self her
hiccup
“She HERSELF is having hiccups.”
-
Assuming an active impersonal construction containing pro, I argue that:
-
8
In several intransitive verbs, the subject is accusative and may be controlled by
k’aye, an emphatic reflexive.
I argue that this is evidence that T does not bear an [EPP] feature, otherwise the
subject of (18) would be in [Spec, TP] and nominative.
There is no [EPP] feature in T, and pro may remain in-situ.
Orthography and glossing modified from O’Connor (1992).
5
-
ApplP is used to check Oblique Case and is complement to vP.
In short distance binding, pro is unable to c-command logophoric pronouns.
(21) tiyaʔ
diyi
ma:dal
čima -ka
-ya
NCBR.Obl.pl next
her
sit
-caus -ya
“She was made to sit next to them.”
(19) * tiyaʔ
hayunam čaban -ʔa
NCBR.Obl.pl dog.spec kill
-ya
* “They killed their own dog.”
The causative marker adds the
external argument pro in
[Spec, VoiceP] directly above.
It is not clear whether hayunam
moves up, but it is the internal
argument and must be generated
within √P. Regardless, oblique
case must be checked at ApplP.
C-command of the
logophoric pronoun
is not possible.
(20) *tiʔ
dakʰe čano matu -ya
NCBR.Obl.sg about story-tell -ya
* “A story was told about herself.” (by her)
The so-called ‘Condition
E’ (Minkoff 2004) that
anaphors (i.e. logophors)
may be linked to
antecedents outside of
their domain would not
hold in this example.
Notice that tiyaʔ diyi
must appear before
ma:dal. Without
ApplP, the only
other position would
have been within √P,
giving us the
incorrect word order.
Scrambling does not
apply here – the
pronoun must appear
to the left of diyi.
Again, c-command of
the logophoric pronoun
is not possible.
6
(22) tiʔ
diyi čima -ka -kʰe
ma:dal mayuʔ -ya
NCBR.Obl.sg next sit
-caus –fut her
flirt
-ya
“(Somebody) flirted with her to get her to live with/sit next to him.”
In this case PRO ccommands the logophoric
pronoun, but pro can
sufficiently do so as well.
(23) man tiyaʔ
hayunam diyi wan -da
ma:dal
she
NCBR.Obl.pl dog.spec next walk -Acomp her
“While she was walking with theirj dog, theyj found her.”
mako -ya
found -ya
It seems that binding
may not be sensitive to
verb phrases as phases.
7
6 Conclusion
While I have barely scratched the surface, evidence suggests that -ya is a marker of
impersonal active clauses rather than impersonal passive clauses. Under this
assumption, pro (which bears a [+human] feature) is realized as the external
argument for transitive verbs and helps explain long vs. short distance binding of
logophoric pronouns. Furthermore I argue that in Northern Pomo, finite-T does not
bear an [EPP] feature. Finally, I provide further support for Harley’s analysis that
verb phrases are made up of three parts, VoiceP, vP and VP, and motivate the use of
ApplP which is complement to vP in order to (a) illustrate possible binding of
logophoric anaphors, and (b) satisfy the word order in (20).
7 Future studies
-
The interaction of reflexives, casuatives and passives.
The structure of impersonal intransitives vs. transitives.
A reanalysis of emphatic reflexives and control phenomena.
References
Maling, J. and S. Sigurjónsdóttir. 2002. “The New Impersonal Construction in
Icelandic.” Journal of Comparative Germanic Linguistics 5: 97-142.
Maling, J. 2015. “A syntactic Rubin's Vase: the inherent ambiguity of nonpromotional passives and unspecified subject constructions”. Presented at: LSA
2015 Annual Meeting. Portland, Oregon.
McCloskey, J. 2007. “The Grammar of Autonomy in Irish.” Natural Language and
Linguistic Theory 25: 825-857.
Minkoff, S. 2004. “Consciousness, Backward Coreference, and Logoricity.”
Linguistic Inquiry 35: 485-494.
O’Connor, M. C. 1992. Topics of Northern Pomo grammar. Outstanding
Dissertations in Linguistics. Garland Press.
O’Connor, M. C. and J. Maling. 2014. “Non-promotional passives and unspecified
subject constructions” In Perspectives on Linguistic Structure and Context: Studies
in honor of Knud Lambrecht, ed. by Stacey Bourns and Lindsy Myers, 17-38.
Amsterdam: John Benjamins. Roberts.
Wurmbrand, S., 2003. Infinitives: Restructuring and Clause Structure. Mouton de
Gruyer, Berlin.
Zaenen, A., J. Maling, H. Thráinsson. 1985. “Case and grammatical functions: the
Icelandic passive.” Natural Language and Linguistic Theory 3: 441-83.
Ackerma, P. and A. Neeleman. 1998. “Optimal questions.” Natural Language and
Linguistic Theory 16: 443-490.
Baker, M. 1985. “The mirror principle and morphsyntactic explanation.” Linguistic
Inquiry 23 (4), 517-555.
Burzio, L. 1986. Italian Syntax. Dordrecht, Holland: D. Reidel Publishing Company.
Collins, C. 2005. “A Smuggling Approach to the passive in English.” Syntax 8 (2),
81-120.
Deal, A. and Catherine O'Connor. 2010. In S. Lima (ed.), Proceedings of SULA 5,
pp 173-188. Amherst: GLSA.
Jaeggli, O.1986. “Passive.” Linguistic Inquiry 17: 587-622.
Kiparsky, P, 2013. “Towards a Null Theory of the Passive.” Lingua 125: 7-33.
Koontz-Garboden, A. 2007. “States, changes of state, and the Monotonicity
Hypothesis.” Doctoral Dissertation. Stanford University.
Krazter, A. 1996. “Severing the external argument from its verb.” In: Rooryck, J.,
Zaring, L. (Eds.), Phrase Structure, the Lexicon. Springer.
Harley, H. 2013. “External arguments and the Mirror Principle: On the distinctness
of Voice and v.” Lingua 125, 34-57.
Harley, H. 2014a. “Aggressive passives in Hiaki.” Presented at NELS 45. Boston,
MA.
Harley, H. 2014b. “On the Identity of Roots.” Theoretical Linguistics 40: 225-276.
Maling, J. 1993. “Unpassives of Unaccusatives.” Unpublished ms., Brandeis
University.
8