| EDITORS: Johns, Alana; Massam, Diane; Ndayiragije, Juvenal
SUBTITLE: Emerging Issues
SERIES: Studies in Natural Language and Linguistic Theory
Patrycja Jablonska, Institute of English Studies, Wroclaw University
This collection is a result of a workshop on theoretical issues concerning
ergativity held at the University of Toronto in 2002. It consists of four parts
(I. The Cases; II Splits; III Antipassive; IV The Range of Ergativity). What
follows is a short summary of the articles included, followed by an evaluation
section. Where relevant to a particular chapter, evaluation immediately follows
summary within square brackets. Whenever the year is not provided, it should be
understood that the present volume is being referred to.
Anand and Nevins (A&N) (''The locus of Ergative case assignment: evidence from
scope'') argue against the ergative-as-nominative hypothesis by bringing in novel
evidence from scope freezing effects. To wit, nominative indefinite subjects in
Hindi allow both wide and narrow scope with respect to a universally quantified
object, while for ergative subjects the latter option is excluded. The authors
tie this fact to the impossibility of an ergative subject to reconstruct below
the QR-ed object, where the assumption is that the factor enabling
reconstruction is AGREE(ment) with T. As ergative subjects are merged lower than
T (in Spec,vP) and assigned inherent case by (defective) v, they will not
constitute the relevant Goal for T's phi-feature valuation. Furthermore, the
authors dissociate ACC from 'objective' (-ko) Case by having the two assigned by
different projections (v and EncP respectively). This, in turn, stems from the
possibility to switch OFF the parameter of obligatory Case assignment by v or T.
Massam (''Neither Absolutive nor Ergative is Nominative or Accusative'') presents
problems associated with equating any of the cases of Ergative/Absolutive
(henceforth E/A) systems with any of the cases of Nominative/Accusative
(henceforth N/A) systems from the Niuean perspective. First, Massam argues that
Ergative behaves as Nominative in being the highest Case and undergoing
processes like relativization, quantifier float, etc. (referring to Seiter
(1980)), but it also displays inherent Case properties, namely it is tied to
agentivity/volitionality and transitivity. Second, Absolutive is argued to be
internal (or akin more to object Case in N/A languages) on the basis of the
following facts: (i) being selected by a preposition; (ii) double ABS occurring
in applicative constructions (iii) the presence of ABS in nonfinite
'subjunctive' ke clauses.
[Generally speaking, it seems that a lot of discussion relating to the typology
of approaches stems from the fact that the notions 'external' and 'internal'
case are not precisely defined. In particular, in systems where AGREE can value
Case on a DP in situ, a potential problem arises: is ''external/internal''
supposed to refer to the position of the functional head that values Case or the
position of the DP? Moreover, even if the former definition is assumed, which
heads count as 'external'? ]
Bobaljik and Branigan (B&B) (''Eccentric Agreement and Multiple Case checking'')
examine a curious piece of data from Chukchi, where in certain ''prohibited''
combinations where Object outranks the Subject on the person hierarchy, the
so-called ''spurious antipassive'' is inserted. The effect of this antipassive,
however, is unusual in the sense that the verb remains syntactically transitive,
but it shows intransitive agreement with Subject only. B&B take both ERG and ABS
to be structural cases assigned by T (due to the deficient nature of the light
verb, cf. Nash, 1995). An offending combination of features on one head is taken
to result in the post-syntactic deletion of object features, yielding
intransitive agreement. Furthermore, following the work in the spirit of copy
theory of movement (e.g. Boskovic, 2002) the authors argue that the late
deletion the object's feature bundle results in activation of the lower copy.
Since the object staying within VP is the configuration relevant for
antipassives, the 'spurious' antipassive morpheme gets inserted.
The contribution by Otsuka (''Syntactic ergativity in Tongan'') focuses on one
diagnostic of the so-called 'syntactic ergativity', namely relativization. The
claim is that 'syntactic ergativity' is derivative with respect to morphological
ergativity, i.e. accessibility for relativization correlates with markedness of
Case in accordance with the scale: Unmarked Case (NOM/ABS) > Marked Case
(ACC/ERG) > Oblique. The author follows an analysis in Suner (1998) and derives
the different patterns of relativization (i.e. resumptive pronouns vs gap
licensing) through varying the presence and the strength of [+pron(ominal)]
feature on C. A strong [+pron] feature attracts a relative pronoun, yielding a
gap, whereas a weak feature licenses in-situ spell out of the relative pronouns
(=resumptive pronoun). Furthermore, C is associated with a Case feature:
+Active, -Active or Unspecified. The implementation of the particular empirical
observation is that a strong [pron] feature is always specified for Active case
[It is a little surprising that Otsuka, analyzing a phenomenon which seems to
suggest equating NOM with ABS, still decides on the analysis where NOM=ERG and
ACC=ABS (cf. Bobaljik 1993) . It seems many of the technical problems could be
avoided had this assumption been dropped.]
Ura (''Aspectually conditioned split-ergativity'') aims to derive tense/aspect
conditioned ergativity splits. Ura proposes two parameters: (i) the light verb
can(not) assign ERG case in situ to the argument in its Specifier; (ii) checking
of ERG Case in situ results or not in 'inert Case'. The difference between deep
and shallow ergativity is taken to follow from the latter parameter: an 'inert
Case' renders the ergative argument invisible for the purposes of feature
checking and the absolutive object needs to move to Spec, InflP for EPP reasons,
which yields deep ergativity. If, on the other hand, no 'inert case' results,
the ergative argument counts as the first accessible Goal and checks EPP on
Infl, yielding shallow ergativity.
[Ura follows a bulk of literature (e.g. Borer 1994, van Hout 1998) where the
telicity/boundedness/delimitation of a predicate requires movement of a DP to
the Specifier of an aspectual projection. Ura extends this type of proposal to
outer/grammatical aspect by saying that the perfective aspect implies the
presence of an aspectual feature on v that the Subject is forced to check off.
To the extent, however, that there is a certain relation between the
interpretation of a DP and the telicity of the clause (as reflected, e.g., in
mereological or compositional approaches to aspect, cf. Krifka, 1992 and
Verkuyl, 1999), it is not clear how grammatical aspect should relate to
arguments. E.g. in N/A languages some perfective verbs don't even require the
presence of an object, as e.g. so-called po-fective verbs in Slavic languages.]
Two senses of the notion 'split' are the topic of Legate's article ('Split
absolutive'). Firstly, the article argues for a split source of absolutive case
in Warlpiri: in intransitive clauses it is assigned by a vP-external head like
T, whereas in transitive clauses it is equal to accusative case (with the
morphological default zero spell out). Secondly, the well know split between
dependent marking (E/A) and agreement (N/A) is argued to follow from the fact
that the morphological spell out of subject agreement can be sensitive to the
morphological case on the Subject. Three types of evidence for the split nature
of ABS are evoked: (i) diagnostics that treat both Absolutive intransitive
subjects and Ergative subjects as a natural class (i.e. agreement, switch
reference system and control); (ii) different distributional properties of
Absolutive Subjects and Absolutive Objects in non-finite clauses; (iii)
coocurrence of two Absolutive arguments with 1,2 person pronouns.
A particular domain of 'split ergativity', namely the Progressive ari in Basque,
is discussed by Laka (''Deriving Split Ergativity in the Progressive''). A rather
rare feature of this account is treating the choice of case system as primitive.
Thus, Laka proposes that the (western) Basque progressive ari is a lexical verb
selecting a postpositional phrase (which in turn selects a nominalised verb
phrase). This structure is biclausal, the higher verb is unaccusative and
assigns Absolutive case to its only argument (i.e. the Subject receiving Theme
interpretation). In eastern varieties of Basque, on the other hand, the
Progressive verbs have been grammaticalized and are now occupying a functional
aspectual head, taking an imperfective verb as its complement (i.e. a
[It seems that an implicit assumption in the analysis is that case assignment is
necessarily tied to Theta roles (i.e. ABS to themes and ERG to agents), which
makes even ABS an inherent case of sorts. It is not clear to me, however, in
what structural sense the provided structure (e.g. (9)) is unaccusative, as the
argument is generated VP-externally. One wonders also how the less-agentive
interpretation of the Subject of the Progressive is to be reconciled with the
fact that cross-linguistically it is Progressive forms of 'unaccusative' verbs
that pass unergativity diagnostics (cf. e.g. Zaenen, 1993 ).]
Wiltschko (''On ''Ergativity'' in Halkomelem Salish'') argues for an epiphenomenal
nature of 'ergativity' based on two types of split in Halkomelem: (i) Person
split (with 3 Person arguments agreement seems to be E/A; with local Person
arguments it is N/A); (ii) clause type split (E/A pattern of agreement in
indicative clauses with 3 Person subject, N/A in subjunctive (negated) clauses
with all persons). The lack of agreement suffix for Intransitive Subjects is
derived through parametrizing the way external arguments might be introduced:
Subjects of unergatives are argued to be introduced in the lexicon (as opposed
to Subjects of transitives). Due to this fact v is missing in unergatives and
the lack of agreement (with absent v) follows. Furthermore, it is proposed that
the splits arise from the fact that 1,2 Person agreement is located in C,
possessive and subjunctive agreement in I and ergative agreement in v.
[The account proposing a lexical derivation for unergatives, however, seems to
raise a question about the extent to which this type of strategy for deriving
unergatives is available cross-linguistically, with a concomitant lack of
syntactic differences between unergatives and unaccusatives.]
Carnie and Cash Cash's article (''Tree-Geometric Relational Hierarchies'') is an
attempt to elucidate the four way case system of Nuumiipuutimt (known also as
Nez Perce, Sahaptian). In this language transitive clauses may appear in two
different patterns: (i) both arguments with zero morphological realization of
case; (ii) Subject in the so-called ergative case (-nm/nim) and Object in the
so-called objective case –ne. The authors argue that there are certain
interpretative differences between the two patterns and that pattern (i) is
essentially intransitive. As in many other accounts, they propose to locate the
structural difference between pattern (i) and (ii) in the various ''flavors'' of
v: intransitive v lacks case features and the subject needs to raise to Spec,TP
for Case (and EPP reasons). Transitive v, on the other hand, has both lexical
Ergative case and structural Acc (i.e. –ne) case features and hence the object
tucks in to the internal specifier of v for Case reasons, and the Subject raises
only for EPP reasons. The authors discuss and reject Diesing's (1992) Mapping
Principle and adopt instead Carlson's recasting of this hypothesis, where the
driving force behind movement is really the semantics of NP types and their
compatibility with event semantics.
Spreng (''Antipassive morphology and case assignment in Inuktitut'') argues (i)
against a non-overt allomorph of the Inuktitut antipassive –si- (as e.g. in
Bok-Bennema,1991, Bittner, 1987); (ii) against a nominal incorporation analysis
(cf. Baker, 1988), and (iii) for a non-uniform analysis of structures with –mik
marked arguments. Thus, she adopts a three-way case system of Woolford (2004)
where Absolutive case is structural case assigned by T, Ergative is inherent
case assigned to Agents by v, and –mik is (i) a structural accusative with
causativized unaccusatives and inherently transitive verbs; (ii) lexical case
assigned by a lexical head with unergative verbs.
[The latter context (i.e. unergative verbs) crucially relies on a root
incorporation into v analysis of unergatives (a la Hale and Keyser (1993)). The
following question arises. If structurally assigned –mik is essentially
equivalent to Accusative case in N/A languages, why should it induce an
indefinite interpretation of the object (as seems to be evident in all the
relevant examples e.g. (10b), (12a), (13a) except (11b)). Admittedly,
differential object marking occurs also in N/A languages (e.g. in Turkish), but
the direction of semantic change is usually the reverse: it is specificity that
an overt case marker induces.]
Ndayiragije (`The Ergativity Parameter: A view from Antipassive'') draws an
analogy between antipassives in ergative languages and reciprocals in certain
Bantu languages. The author argues that Chichewa and Kirundi reciprocal –an- is
a little v–head ''underspecified for a [Null Case]''. As the claim concerning
ergativity parameter is that T and v lack structural Case in 'pure ergative
languages', the parallel between antipassives and reciprocals follows: in both
cases the lack of structural Case induces the presence of PRO in the object
position. Ndayiragije observes the split into Chichewa-type reciprocals and
Kirundi-type ones, where the former are restricted to transitive verbs only and
the latter receive a comitative reading when applying to intransitives. It is
argued that the split is due to different featural specification of the relevant
morpheme: Chichewa –an- is Theta-defective (and therefore needs an object to
discharge its Null Case feature); Kirundi –an- is Theta-ambivalent (i.e. can (i)
either introduce its own argument interpreted generically or reciprocally or
(ii) not introduce an argument at all requiring a PRO object and thus giving an
impression of arity reduction.
John's article (''Ergativity and change in Inuktitut'') adds an important
diachronic dimension to the volume. Similarly to Spreng (2001, this volume), she
argues that the antipassive –si- in Inuktitut is located in v. The ambiguity
between inceptive and antipassive reading of the morpheme in question is
explained by proposing that in both cases it contributes interpretable
indefinite quantity features, but is merged outside vP in the inceptive use, and
in vP in the antipassive. The –mik case is taken to be a syntactic default case
possessing uninterpretable features. The difference between Western dialects
(where antipassive is more marked) and Eastern dialects (where it is widespread)
is taken to stem from the fact that in the former –mik lacks a Match property
(i.e. it is not associated with a specific case licenser, and hence is used in a
lot of other constructions) whereas in the latter –mik requires a little v as a
Case licensing head. Thus, Johns argues, -mik in Eastern dialects (as in
Labrador Inuttut) is more akin to (structural) accusative case (cf. also
THE RANGE OF ERGATIVITY
Paul and Travis (P&T) (''Ergativity in Austronesian languages'') test the limits
of the ergative hypothesis for Malagasy. They review a variety of diagnostics
which ergativity helps explain (e.g. binding, control, extraction and
imperatives), as well as those that seem problematic for the ergativity
hypothesis (object status, lack of weak cross-over, optionality of Actor). An
important comparative upshot of the discussion is the proposal which the authors
label as 'ergativity continuum': Tagalog is more ergative than Malagasy (because
of e.g. an indefinite reading of the Theme in Antipassive), which in turn is
more ergative than Bahasa Indonesian (because of control contexts). On the other
hand, from the language internal perspective, there exist constructions which
are ergativity or accusativity domains (as the two types of 'passives' in Bahasa
Indonesian discussed first by Chung (1976).
[One issue that comes to mind is that the language-internal variation argued for
here opens up a possibility that even for Tagalog the Agent Topic construction
might be a N/A domain rather than an antipassive construction. It seems that a
part of the problems related, e.g., to the status of the object was problematic
only for the 'antipassive subhypothesis'. ]
Tsedryk (''Split verbs as a source of morphological ergativity'') investigates
another ergativity domain within otherwise canonically accusative languages
(i.e. Dative psych-verbs in Russian and participial –n/t- clauses in North
Russian). The gist of the proposal is that in ergative languages/constructions
the light verb is split into two projections: v1P hosting Theta-features and v2P
hosting phi-features. Since the conditions for AGREE are defined in terms of
c-command, and since incorporation of v1 into v2 destroys the c-command relation
between v1 and the object, only AGREE with the Subject is possible, yielding
ERG. In accusative languages/constructions, on the other hand, both Theta- and
phi-features coalesce on one head and v enters AGREE with the object, which
yields Acc case. Under this set of assumptions the author derives various
case/agreement patterns in participial clauses in North Russian by postulating a
deficient participial v and altering its feature make-up.
The book is an important and valuable resource due to two factors: (i) its
theoretical import concerning issues related to ergativity, and (ii) data from
many endangered languages, often coming from the authors' own fieldwork. The
somewhat narrow theoretical framework (all articles written within some version
of the Minimalist Program (Chomsky, 2000, 2001)) in fact constitutes a virtue,
as the accounts are easily comparable. On the other hand, it also exposes the
inadequacy of certain types of explanation being pursued. In what follows I
concentrate on several subjectively chosen ''emerging issues''.
In many papers (e.g. Legate, A&N, B&B, Tsedryk) there occurs a projection, which
is ''deficient'' with respect to some feature (most often it is case assignment).
E.g., Legate is forced to assume that a 'deficient T' of sorts is present in the
Numeration in a very specific set of circumstances (i.e. in a transitive clause,
but only in case the subject is everything else but 1,2 pronoun). Otherwise,
uninterpretable features on T should result in a crashing derivation. The
question arises whether such tinkering with the Numeration is conceptually
desirable (as it virtually amounts to (long-distance) selection) or
computationally plausible. A similar issue arises also in A&N's article, where
Hindi v-def occurring in perfective contexts lacks the ability to assign ACC.
From the point of view of (long-distance) selection anything can happen. Thus,
one expects also to see a language where an imperfective context would require
the presence of a defective v. This, however, never seems to be the case
An interesting upshot of Legate's analysis is the fact that certain syntactic
properties of arguments with a particular case marker are essentially taken to
be due to a language-specific lexical accident, i.e. the lack of ACC case suffix
in Warlpiri. Thus, instead of ACC, a 'morphological default' (i.e. ABS) is
inserted. It seems to me to be more questionable whether such an account could
rely on the 'morphological default'. If Niuean were to also be analyzed in these
terms, as the author suggests in fn. 6, then probably no 'morphological default'
can be invoked, as the absolutive case marker is not phonologically zero in
Niuean. In general, whenever a case marker occurs in various contexts, different
types of 'defaults' are conjured up (e.g. 'syntactic default' in Johns,
'morphological default' in Tsedryk). The question arises, however, whether the
specific case labels should not be treated as possibly multiply ambiguous.
For instance, one of Massam's arguments for 'internal' ABS in Niuean hinges on
equating ABS assigned in ''bare'' arguments and the one selected by particular
prepositions. Yet, to the extent that ABS is taken by Massam to be a structural
case assigned in a particular configuration, it is not clear how two ABSs can be
present in a clearly monoclausal structure in (1) (Massam's (13)).
(1) Ne kai e Sione e tau talo aki e huki.
PST eat ERGP Sione ABSC PL taro with ABSC fork
'Sione ate the taros with a fork.'
For this reason it seems to me that the possibility that the Niuean ABS marker
is multiply ambiguous should be seriously considered.
In conclusion, cross-linguistic discussion involving traditional Case labels
like NOM, ERG, ABS, ACC misses the point if these Case markers can be
multifunctional or ambiguous between different syntactic contexts. A case marker
X (referred to as e.g. ABS) in a language x can occur in contexts A, B and C. As
far as I can see; however, there is nothing that would guarantee that a case
marker Y (also referred to as ABS) in a language y should share any of the
contexts with X.
Universality of diagnostics
It becomes conspicuous, especially in Wiltschko's contribution, that the
standard definition of ''ergative behavior'' (i.e. treating S and O as a natural
class, as opposed to A) is quite inadequate.
As far as I can see, it is often unclear whether a particular behavior with
respect to a certain test should be treated as indicative of 'ergativity' or
not. Let's consider the famous 'relativization' test, which is usually taken to
be a diagnostic of 'syntactic ergativity'. If the claim in Otsuka to the effect
that relativization is always tied to the morphological markedness of case is
true, then in no language will an ergative or accusative DP be more accessible
for relativization than an absolutive or nominative DP (although both might be).
In this sense E/A and N/A languages behave identically. Otsuka implements it in
terms of strong and weak features on C and the specification for +/-Active Case
or lack thereof. Thus, e.g. in Niuean, where both Unmarked ABS and Marked ERG
arguments license a gap (as opposed to resumption), C is taken to have a strong
feature with Unspecified Case (so that it can agree with a relative pronoun
regardless of its Case). In fn. 15, however, the author notices that indirect
objects and obliques in Niuean require resumption. It is not clear to me how
this distinction should be implemented in the proposed system. What's more, as
observed in Taraldsen (2006), an even more fine-grained distinction arises in
Icelandic, where Dative obliques also license a gap (i.e. do not require
resumption). To sum up, it seems that languages simply have different cut-off
points along the Case markedness scale where licensing the gap stops, with the
relevant (yet possibly separate) question being why the resumption strategy is
not available in Dyirbal (cf. Dixon, 1994). Yet, there is nothing 'ergative',
the more so nothing 'deeply ergative' about this particular diagnostic.
A further problem is involved in the data complexity. E.g., P&T, when
substantiating their hypothesis concerning 'ergativity continuum', invoke the
control argument. As Arka and Manning (1998) claim that control in ergative
languages is sensitive to the Agent Theta role, the fact that the controller in
Bahasa Indonesian is 'determined by structure/case' makes it more of a N/A
language. I take it that the difference between Malagasy and Bahasa Indonesian
is the existence of object control in the former (ex. (22)). If that was the
intention, it seems that here control by the Absolutive argument is taken as
'ergative behavior'. But canonically N/A languages like English obviously also
have object control with the relevant verb 'ask'. It seems that these are not
the control contexts sensitive to the Theta role that Arka and Manning (1998)
analyzed (although I could not consult the source).
A final complication involves the question related to domains of 'ergativity'
within a N/A language. The discussion in P&T involves two 'passives': one of
them being an English-type passive and the other being akin to Theme Topic in
Malagasy. The former serves as an example of a more N/A behaviour (i.e. similar
to English). There remains, however, a possibility that the passive construction
is precisely the domain of 'ergativity' in English, in which case comparison
with the relevant diagnostics in the passive in English is not very telling.
It seems tempting to my mind at least to turn the question on its head, i.e. to
assume that the results of the diagnostics are identical, irrespective of the
case system, once it is understood what a particular diagnostic is sensitive to.
One such attempt for relativization is Otsuka, another one for agreement is
Bobaljik (2007), where it is argued that agreement is universally determined by
the hierarchy of case markedness. Admittedly, Halkomelem Salish (cf. Wiltschko)
would constitute a counterexample, as it displays agreement with Ergative A, but
not with 3 Person Absolutive S.
Arka, I.W. and Manning, Ch. (1998) Voice and grammatical relations in
Indonesian: a new perspective. _Proceedings of the 1998 International Lexical
Functional Grammar Conference_.
Baker, M. (1988) _Incorporation: A Theory of Grammatical Function Changing_.
Chicago: University of Chicago Press.
Bittner, M. (1987) ''On the semantics of the Greenlandic antipassive and related
constructions.'' _IJAL_ 53:2, 194-231.
Bobaljik, J. (1993) ''On ergativity and ergative unergatives.'' _MITWPL_ 19, 45-88.
Bobaljik, J. (2007) ''Where's Phi? Agreement as a post-syntactic operation.'' In
Adger, D. and Bejar, S. and Harbour, D (eds.) _Phi features across interfaces
and modules_. Oxford: OUP.
Bok-Bennema, R. (1991) _Case and agreement in Inuit. Studies in Generative
Grammar_ 38. Berlin/New York: Foris Publications.
Borer, H. (1994) ''The projection of arguments''. _University of Massachusetts
Occasional Papers_ 17, 19-47.
Boskovic, Z. (2002) ''On multiple wh-fronting.'' _Linguistic Inquiry_ 33, 351-383.
Chomsky, N. (2000) ''Minimalist inquiries: The framework.'' In Martin, R. and
Michaels, D. and Uriagereka, J. (eds.) _Step by Step: Essays on Minimalist
Syntax in Honor of Howard Lasnik_, 89-155. Cambridge, MA.: MIT Press.
Chomsky, N (2001) ''Derivation by phase''. In Kenstowicz, M. (ed.) _Ken Hale: A
Life in Language_, 1-52. Cambridge, MA.: MIT Press.
Chung, S. (1976) ''On the subject of two passives in Indonesian''. In Li, C. (ed.)
_Subject and Topic_. New York: Academic Press.
Hout, van A. (1998) _Event Semantics of Verb Frame Alternations_. New York: Garland.
Krifka, M. (1992) ''Thematic relations as links between nominal reference and
temporal constitution.'' In Sag, I. A. and Szabolcsi, A. (eds.) _Lexical
Matters_. Stanford: CSLI.
Nash, L. (1995) _Portee argumentale et marquage casuel dans les langues SOV et
dans les langues ergatives: l'example du georgien_. These de doctorat,
Universite Paris 8.
Spreng, B. (2001) ''Little v in Inuktitut: Antipassive revisited.'' _Linguistica
Suner, M. (1998) ''Resumptive restrictive relatives: A cross-linguistic
perspective.'' _Language_ 74, 335-364.
Taraldsen, K. T. (2006) ''Case and resumptive clitics in Czech.'' abstract for the
talk at _FDSL_ 6.5, Nova Gorica.
Woolford, E. (2004) ''Lexical case, inherent case and argument structure.'' Ms.,
Verkuyl, H. (1999) _Aspectual Issues: Structuring Time and Quantity_. Stanford,
Ca: CSLI Publications.
Zaenen, A. (1993) ''Unaccusativity in Dutch: intergrating syntax and lexical
semantics.'' In Pustejovsky, J. (ed.) _Semantics and the Lexicon_. Dordrecht: Kluwer.
ABOUT THE REVIEWER
Patrycja Jablonska is an Assistant Professor at the English Institute,
University of Wroclaw. She works on the structure of the thematic domain, issues
related to argument structure, argument encoding, the status of semantic
features like 'animacy' in linguistic theory, aspectual issues.