Volume 35, Issue 3 - GLOW Issue

In this paper, I argue that Spanish prenominal and postnominal possessives target different external merge positions focusing on alienable possessive constructions. The analysis is developed alongside a proposal on the organization of DPs, according to which articles are merged as a DP-internal category between the domains assigned to direct and indirect modifiers. Prenominal possessives are determiners reanalyzed from direct modification adjectives whereas postnominal possessives are indirect modification adjectives that arise as predicates of reduced relative clauses. This analysis provides a principled explanation on the behavior of Spanish possessives that is lacking in the generalized idea that they are pronouns with a unique merge position. Arguments are also presented showing that syntax-driven phonological restrictions condition the derivation of DPs. The account successfully derives the core properties of word order variation and related issues within possessive constructions.

Modern generative linguistic theory furnishes a variety of general principles that appear to be at work in the grammar of all the world’s languages. One of the most basic and uncontroversial of these principles is that Agree/Move operates according to the constraint Attract Closest , which dictates that the closest suitable goal must be the target for the relevant operation (Rizzi 1990; Chomsky 1995, 2000; Richards 1998). The Polynesian language Niuean (Tongic subgroup, predicate initial word order, ergative-absolutive case system) presents a well known challenge to the universality of {Attract Closest}. The challenge manifests in a variety of distinct constructions in Niuean, but the best known case involves an operation first documented by Seiter (1980), which he terms “raising.” Specifically, Niuean raising appears to license an A-type dependency between the subject position of the matrix clause and the object position of an immediately embedded clause. This is illustrated in (1), where the semantic object of the embedded subjunctive clause, Sione , appears as the syntactic subject of the matrix predicate maeke . (1) To maeke a Sione 1$_{1}$ [ke lagomatai he ekekafo t1$t_{1}$ ] . fut possible abs Sione sbj help erg doctor ‘It’s possible the doctor can help Sione.’ (lit.: Sione is possible that the doctor help [him]) Granting that the filler-gap dependency in (1) is A-type, this is both a clear violation of {Attract closest} (Rizzi 1992; Chomsky 1995; Richards 1998) and a typological anomaly. Our aim in this paper is to argue that such apparent violations of {Attract Closest} are only that. Specifically, we show first that the challenge inherent in Seiter’s raising construction is pervasive throughout the language: in general, objects are accessible to syntactic operations even if the intervening clause-mate subject is also a licit target. In other words, Niuean clause-mate subjects and objects are equally accessible to syntactic operations. Then, we argue that this typologically uncommon equal-accessibility follows from the convergence of several otherwise independently attested operations: (i) a configurational system of case licensing, with a v$v$P as the case computation domain; (ii) obligatory object shift to Spec(v$v$P); (iii) an EPP on T triggering V/VP-raising rather than DP externalization. The resulting basic clause structure is then as below, so that Niuean adheres to standard locality constraints. (2)

The goal of this paper is to provide novel theoretical and empirical evidence that the null subjects traditionally labelled as pro and pro , rather than being inherently distinct, are manifestations, differentiated in the course of the derivation, of what is underlyingly a single underspecified nominal pro-form, which we will call UPro. Included under this UPro are pro , oc pro and also the various types of ‘non-obligatory control’ ( noc ) pro , including arbitrary pro ( pro arb ). The interpretive and distributional distinctions lurking behind these labels result from how UPro interacts with its structural environment and language-specific rules of morpho-phonological realization. Specifically, oc pro labels a rather specific interpretation that arises in embedding contexts where a syntactic oc relationship with an antecedent can be established. Different types of pro and noc pro , on the other hand, involve ‘control’ by (typically) silent representations of discourse-contextual elements in the clausal left periphery. Finally, pro arb$_{arb}$ arguably involves the failure to establish a referential dependence, which we will formalize in terms of a failure to Agree in the sense of Preminger (2014). Crucial evidence motivating the approach proposed here will be adduced from Sundaresan’s (2014) “Finiteness pro-drop Generalisation”, which reveals an otherwise unexpected complementarity of oc pro and pro .

This paper addresses two restrictions regarding agreement with nominative arguments in Icelandic DAT-NOM constructions. The first is the reported asymmetry in intervention effects in mono-clausal versus bi-clausal environments. The second regards the well-known Person Restriction that prohibits agreement with non-3rd person arguments. It is argued that both of these phenomena can be viewed as instances of cumulative constraint interaction, where less important constraints in the grammar ‘gang up’ to block some higher constraint. In order to account for this, I adopt a model of syntax with both weighted constraints and serial optimization that is known as Serial Harmonic Grammar in the phonological literature. It will be demonstrated that such a system can offer a more principled analysis of the construction-specific nature of the aforementioned phenomena.

In their recent paper, Lau, Clark, and Lappin explore the idea that the probability of the occurrence of word strings can form the basis of an adequate theory of grammar (Lau, Jey H., Alexander Clark & 15 Shalom Lappin. 2017. Grammaticality, acceptability, and probability: A prob- abilistic view of linguistic knowledge. Cognitive Science 41(5):1201–1241). To make their case, they present the results of correlating the output of several probabilistic models trained solely on naturally occurring sentences with the gradient acceptability judgments that humans report for ungrammatical sentences derived from roundtrip machine translation errors. In this paper, we first explore the logic of the Lau et al. argument, both in terms of the choice of evaluation metric (gradient acceptability), and in the choice of test data set (machine translation errors on random sentences from a corpus). We then present our own series of studies intended to allow for a better comparison between LCL’s models and existing grammatical theories. We evaluate two of LCL’s probabilistic models (trigrams and recurrent neural network) against three data sets (taken from journal articles, a textbook, and Chomsky’s famous colorless-green-ideas sentence), using three evaluation metrics (LCL’s gradience metric, a categorical version of the metric, and the experimental-logic metric used in the syntax literature). Our results suggest there are very real, measurable cost-benefit tradeoffs inherent in LCL’s models across the three evaluation metrics. The gain in explanation of gradience (between 13% and 31% of gradience) is offset by losses in the other two metrics: a 43%-49% loss in coverage based on a categorical metric of explaining acceptability, and a loss of 12%-35% in explaining experimentally-defined phenomena. This suggests that anyone wishing to pursue LCL’s models as competitors with existing syntactic theories must either be satisfied with this tradeoff, or modify the models to capture the phenomena that are not currently captured.