Belk, ZCE; (2017) Attributes of Attribution. Doctoral thesis , UCL (University College London).

Preview

Text Belk.Attributes of Attribution (submitted 20.04.2017).pdf Download (1MB) | Preview

Abstract

Attributive adjectives have a number of properties that distinguish them from predicative adjectives and other modifiers of the noun. For example, attributives are subject to ordering restrictions that do not apply to other types of nominal modifier, and they exhibit scope interactions unlike predicatives or relative clauses. This thesis argues that these properties are best captured by an analysis in which all attributives share the same relationship with the noun and discusses the ways in which our understanding of a number of phenomena at the edges of attribution must change. One influential theory of adjective ordering restrictions (discussed in Larson 2000a and Cinque 2010, among others) holds that violations of the ordering hierarchy that applies to many attributive adjectives are due to the existence of modifiers that superficially look like attributives but are in fact derived from reduced relative clauses. These derived attributives are merged higher than underived attributives and are unordered with respect to each other. I show that the offending adjectives do not behave syntactically like true relative clauses, whether full or reduced. In addition, while all attributive semantics is asymmetric, true relatives involve symmetric modification. This single-source approach entails a rethinking of some of the effects commonly understood to result from attribution. I will address two such effects, which could be taken as evidence in favour of a derived attributive approach to attribution, and show that they are best analyzed using a homogeneous approach to attribution. In languages where the noun follows its modifiers, the ordering of AP and PP modifiers is free and their scope varies with c-command. In noun-initial languages their order is fixed, with the AP preceding other modifiers, and their scope is ambiguous. This pattern could be taken as evidence for a second source of adnominal modification, if the high position of the AP in noun-final languages is a reduced relative clause. However, I show that both the ordering and scope effects are due to a novel constraint restricting the linear order of attributive and other modifiers. The ordering patterns of AP and PP modifiers are therefore not evidence for the existence of derived attributives. One piece of evidence for the dual-source theory of attribution is that some adjectives have unexpectedly rigid requirements for adjacency and nonintersectivity (for example, in the phrase hard worker). I demonstrate that cases like these are not true attribution but are instead a type of bracketing paradox. I argue that these bracketing paradoxes are derived by movement at LF. This movement (and indeed all movement) is restricted in the type of information that must be retained before and after the operation takes place, but is other- wise free. Therefore, these examples do not provide evidence for two different types of attributive modifier. The proposed analysis of attribution allows for a simplification of adjectival modification, as it does not require a distinction between derived and underived attributive adjectives. The analysis presented in this thesis entails a novel categorization of certain adjectival phenomena, but readily accounts for the empirical intricacies of attribution.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item