The Circle is a weekly forum dedicated to discussion of the descriptive, experimental, and theoretical study of syntax and semantics, featuring presentations of ongoing research by members of the Berkeley Linguistics Department and other departments, as well as discussion of previously published works.
Wednesdays, 3:00-4:30 pm
1303 Dwinelle Hall
University of California, Berkeley
Department of Linguistics
RRR week: TBD
Gosia Szajbel-Keck (UCB): Secondary Predication in West Slavic
In this talk, I will discuss the syntax of secondary predicates (nominal and adjectival) in Polish. Polish is one of the languages, where the use of adjectives and nouns as secondary predicates is highly restricted. This, however, does not mean that there is nothing to say about them. In my analysis of secondary predication, I follow the view that they should be analyzed as small clauses (so called piP in Citko's 2008 analysis of small clauses in Polish). I will show that secondary predicates can be either arguments or adjuncts to the main predicate, they can come with filled or empty pi heads. In addition, both nouns and adjectives must be morphologically marked for gender/number/case, but the case depends on the presence or absence of the pihead, the word class (adjective vs. noun) as well as the case/number/gender features of the controller. In my presentation I will try to explain this complexity by grouping these secondary in three distinct syntactic constructions.
Matthew A. Tucker (UCSC): Two Kinds of Causatives in Maltese
In this talk I will discuss two means of forming causatives in Maltese (Semitic; Republic of Malta). This language allows causatives to be formed synthetically via idiosyncratic stem allomorphy (which I call the morphological causative) as well as periphrastically via use of the causative predicate ġiegħel, 'to make' (which I call the syntactic causative) This talk will examine the morphosyntax of these two constructions and propose that they are both formed in the syntax, with the evidence for this syntactic decomposition coming from possible adverbial attachment sites. The morphological causative will be shown to fit into the proposal by Pylkkänen (2008) wherein causatives contain a single external-argument introducing v head which takes an Applicative projection complement. The syntactic causative, on the other hand, will be shown to require a recursive vP structure involving two heads which introduce two external arguments, a causer and causee. I will then show that the syntactic causative has several important implications for syntactic theorizing, namely: (1) it provides support for a reduced clausal analysis of causative complements akin to the proposals by Wurmbrand (2001) for restructuring predicates, (2) it shows evidence of a polarity position inside the causative complement but no evidence for tense, aspect, or complementizer positions normally seen in clausal complements, and (3) it shows that nominative case assignment and finiteness must be divorced from subject agreement, as the causative complement verb is still inflected for the person, number, and gender of the causee argument. The result is an ECM-like analysis of ġiegħel which can account for all the facts in a unified way. Finally, I will propose, following Marantz (1991) and others, that case assignment in causatives must be disjunctively computed, providing evidence from transitivity alternations and their effect on the case of the causee that a syntactic analysis with no disjunctive morphological case realization is inadequate.
Robyn Orfitelli (UCLA): Argument intervention and A-movement in
Arto Anttila (Stanford) and Jong-Bok Kim (Kyung Hee University): Case and cyclicity in Finnish
We present new evidence for two theoretical claims: (i) case assignment is configurational: the case of an NP depends on the presence of other NPs in the same local domain (Yip, Maling, and Jackendoff 1987, Marantz 1991, Maling 1993, Bobaljik 2008, Baker and Vinokurova 2010); (ii) case assignment is cyclic: case in a complex clause is a function of case in its constituent clauses. The evidence comes from systematic case variation in propositional complements of verbs like ‘say’, ‘think’, ‘want’, and ‘believe’ in Finnish:
Mati-n luul-tiin [ampu-nee-n karhu ~ karhu-n]
Matti-GEN believe-PASS.PAST [shoot-ACT.PERF-GEN bear-NOM ~ GEN]
‘Matti was believed to have shot a bear’
In an experimental study, Itkonen (1976, 1981) noted that in some sentences the judgments are sharp and the case is either NOM or GEN, whereas in other sentences both cases are acceptable to varying degrees. We show that both patterns are predictable if we assume that case assignment is configurational and cyclic. The data provide evidence against the view that case is assigned by nearby functional heads through agreement (Chomsky 1981, Legate 2008, Vainikka and Brattico 2009, Baker and Vinokurova 2010).
Marilola Perez (UCB): Cavite Chabacano (Philippine Creole Spanish)
No meeting - spring break
Marine Vuillermet (UCB): Incorporations into verbs and adjectives in Ese Ejja
Ese Ejja is spoken in the Bolivian and Peruvian lowlands by about 1,500 people. This polysynthetic language displays two types of noun incorporations. The first type of noun incorporation is typical of the Amazonian area: it occurs within verb predicates, is restricted to possessed nouns and has no influence on the valency of the verb predicate. The second type is unusual, as it occurs within predicative adjectives. It results in adjectives, comparable to the English derived adjectives ‘blue-eyed’ or ‘red-haired’. However, the phenomenon is much more productive in Ese Ejja, yielding words like many-sistered or slippery-sloped. More surprisingly, noun phrases and even verbal roots can be incorporated. Following Mithun (1984), I will argue that NI into adjectives is not automatic but serves a specific purpose within the grammar: in this language where attributive adjectives are rare, NI into adjectives corresponds to a strategy to modify NPs.
Nicholas Rolle (UCB): Serial verb constructions in Degema, an Edoid
language of Nigeria
In this talk, I present an analysis of serial verb construction (SVC) in
Degema [Nigeria: Edoid, Benue-Congo, Niger-Congo], primarily looking at the
surface distribution of verbs, objects, and verb phrase clitics, and their
phonological realization. In one SVC pattern, both verbs appear with a
proclitic expressing subject agreement and an enclitic expressing
tense/aspect, referred to as the double-marked pattern. In another, only
the first verb occurs with a proclitic and only the second verb occurs with
an enclitic, referred to as the single-marked pattern. Objects may
intervene between the double-marked pattern, though cannot intervene
between the single-marked pattern. This is summarized below:
1. Double marked: cl=V1=cl (Obj) cl=V2=cl
2. Single marked cl=V1 (*Obj) V2=cl
In this analysis, I present two main claims. The first is that the
difference between these two SVCs is structural: the double-marked pattern
involves two AGR phrases and two ASP phrases, each of which is projected
from a clitic head, whereas the single-marked pattern involves only one AGR
and one ASP phrase. The second claim is these syntactic structures result
in different prosodic structures, which dictate the surface positions of
the objects within the SVC. This work is being conducted under the guiding
question asking under what conditions is the pronunciation of different
copies within a movement chain sensitive to prosodic structure.
Sara Cantor (UCSC) on Sluicing
Bonnie Krejci (Stanford): The Causativization of Middle and Ingestive Verbs: An antireflexivization analysis
Recent typological studies (Dixon & Aikhenvald 2000, Shibatani 2002) have shown that less transitive verbs morphologically causativize more readily than those that are more transitive, delineating four verb classes: unaccusatives, middles and ingestives, unergatives, and transitives. Ingestive verbs, canonically denoting actions of eating and drinking, are unique in sometimes patterning with unaccusatives despite being transitive (compare Amharic k'əllt'ə “melt (intr.)” > >a-k'əll't'ə “melt (tr.)” and bəlla “eat” > a-bəlla “feed”). In some languages, middle verbs, a semantic class canonically denoting self-directed action, pattern with ingestives. That transitive middles are derived from intransitives via causativization at all is unexpected given the typical analysis of middles as reflexives (Kemmer 1993). Middles and ingestives also display analogously unusual meaning differences between their lexical and periphrastic causative variants, indicating that the two classes may undergo the same kind of lexical causativization operation. I motivate an analysis wherein the base forms of these verbs are lexically reflexive with bieventive, causative event structures, and that, in languages where these verbs causativize like unaccusatives, the causativization process delinks the two coidentified arguments rather than adding a causer.
I demonstrate the viability of this analysis based on a detailed account of the lexical entailments of selected English middles and ingestives, in the basic (1a), lexical causative (1b), and periphrastic causative (1c) variants.
1. a. John dressed.
b. Mary dressed John.
c. Mary made John dress.
The lexical and periphrastic forms are not equivalent; only (1c) entails (1a). This suggests that the lexical and periphrastic causatives are derived from the base form via different operations. Additionally, I provide evidence from sublexical modification (Dowty 1979), negation (Koontz-Garboden 2009), and modification with by itself (Chierchia 2004) in support of the position that the base verbs have complex event types, already encoding causation.
This proposal mirrors recent work by Koontz-Garboden (2009), who argues that, in Spanish and other Romance languages that derive anticausatives from causatives, the anticausativization operation is a kind of reflexivization. In cases with the opposite direction of derivation, as in causativizing languages like Marathi, an antireflexivization analysis can capture both the morphological and semantic facts. Such an analysis aligns well with previous proposals to provide a fuller picture of the typology of causativization.
Mark Norris (UCSC), Nominal Concord in Icelandic and Estonian
Nominal concord, a term I use to refer to apparent agreement between a noun head and its modifiers, is a well-known phenomenon in natural language, but it has received fairly little attention, such that a widely accepted analysis of it has remained elusive. In this talk, I propose an analysis of the patterns of concord in Icelandic and Estonian, two genetically unrelated languages with fairly rich systems of concord in (gender), number, and case:
1) all-ar þess-ar litl-ar snigl-ir
all-NOM.M.PL these-NOM.M.PL little-NOM.M.PL snail-NOM.M.PL
'all these little snails' (Icelandic)
2) Kõik nee-d väikse-d teo-d
all.PL.NOM DEM-PL.NOM little-PL.NOM snail-PL.NOM
'all these little snails' (Estonian)
I begin the talk by considering an Agree-based theory of concord, and I discuss some reasons such an approach is unsatisfying at best. I then propose a new analysis composed of two parts: 1) explicit Feature Percolation Principles in the syntax, and 2) postsyntactic Feature Copying.
The analysis treats concord as a mostly morphological phenomenon, and as such, it is less sensitive to syntactic structure than more familiar cases of argument-predicate agreement. I conclude by considering some particularly recalcitrant data from numerals in Estonian, showing that the copying relation in concord must be local.
Melanie Redeye (UCB) on Seneca
Peter Jenks (UCB): The functional architecture of nominal modification in Chinese and Thai (with Shizhe Huang)- abstract Evidence for a unified theory of movement: Moro DPs- abstract
Round Table Discussion on Syntactic Theories
We will be meeting to discuss some recent articles on Construction Grammar and Generative Grammar by Adele Goldberg and David Adger. The three articles can be found here, here, and here
Andrew Garrett (UCB): The Yurok Tense and Aspect System