In this paper I present a minimalist analysis of Semitic Construct State DPs, using mainly data from Makkan Arabic, a spoken variety used in some parts of the western region of Saudi Arabia. I first review the major properties of Construct States and some of the main accounts found in the literature about this structure. I then turn to the analysis I am proposing, which basically consists of two derivational steps: the movement of N to the root of the DP in syntax and a morphological merger operation at the level of Morphological structure. This proposal adds support to recent attempts in the literature to defend head-to-spec movement (Matushansky 2006, and others). I also discuss the differences between Construct State DPs and other Semitic DPs, mainly DPs with overt articles on the noun, and I argue that these differences are due to the fact that the D projected in the each type of DP has some unique features.

This paper uses natural conversational data to examine the use of rhythm in conversation as a resource to generate phonetic and phonological coherence in English lists. This paper does not continue the search for all-pervasive rhythm in language, but considers one particular activity in conversation, that of list making. Data are presented that show that rhythm can indeed be produced and oriented to within the activity of list-making. Rhythm production and orientation is seen to be important to interaction.

This paper shows that Japanese, which does not have a comparative morpheme, can build comparative constructions in a compositional way. Specifically, my focus is on phrasal comparatives, where the complement of yori 'than' is NP. I first discuss the semantics of comparatives, introducing a measure function analysis. Specifically, I show that the direct interpretation of phrasal comparatives is necessary. I then discuss the external and internal syntax of phrasal comparatives, concluding that the yori phrase is selected by an abstract comparative morpheme -er/more. The scopal interaction with other quantificational expressions will also be focused, arguing that comparatives must not take scope over negation.

This paper reports the results of two experiments investigating the effects on speech acoustics and intelligibility of a number of different types of forensically-relevant fabric mouth and face coverings, including the niqāb (full-face Muslim veil), balaclava, and surgical mask. For the perceptual (intelligibility) experiment, subjects were presented with two types of speech stimuli, ‘bimodal’ and ‘unimodal’, and asked to write down what they heard. Four facial guises were used (niqāb, balaclava, surgical mask, no covering). In the bimodal condition (video + audio), subjects saw and heard video recordings of actors reading target words embedded in a standardised carrier sentence. In the unimodal (audio only) condition, subjects heard just the soundtrack of the same video recordings, i.e., no visual image was present. It was found in the perceptual test that the subjects could in all four guise conditions correctly identify target words with a high degree of reliability, and that a small number of confusion types accounted for the majority of the errors. For the second (acoustic) experiment, the objective was to assess the sound transmission loss characteristics of the fabrics from which these and other face coverings are composed. This experiment showed that transmission loss was negligible for all but one of the fabrics, suggesting that speech intelligibility problems created when mouth and face coverings are worn by speakers must derive principally from the reduction in visual information available to listeners and/or from the auditory consequences of interference with speech articulation caused by the face coverings, rather than from transmission loss of the fabrics themselves.

Questions with a wh-object and a universally quantified subject, such as What did everyone buy?, are often reported to lack a pair-list reading in Japanese and Korean, while in English, the pair-list reading is allowed. However, the present paper reports on a challenge to this generalisation posed by robust data from a new experimental study, in which native Korean speakers accepted pair-list readings in Korean. By contrast, native Japanese speakers generally rejected pair-list readings in Japanese, in the same experiment. The paper thus attempts to discover the source of this unexpected interpretation difference between Japanese and Korean. Typological differences are examined between the Korean and Japanese quantifiers nwukwuna/daremo 'everyone'; and a semantic account by Saito (1999) of the lack of pair-list readings in Japanese is outlined. However, the latter is shown to be unable to extend to cover the present Korean data. Finally, a preliminary proposal is sketched, whereby the presence or absence of post-nominal focus marking on the universally-quantified subject determines whether or not pair-list readings are available in Korean and Japanese.

In this paper I investigate differences in the line position of the non-finite verb in the Old English poem Beowulf and Ælfric's rhythmic prose. I show that while in Beowulf the non- finite verb appears in final position in the half-line over 95% of the time when there is an object present, in Ælfric this tendency, while still present, is less strong. I explore how this difference might be related to changes in prosodic structure between the two texts, triggered by word order changes: specifically, the change from head-final to head-initial order in the VP.