It is well known that stress assignment in English noun–noun compounds is non-uniform (compare e.g. left-prominent ópera glasses and right-prominent steel brídge), and recent corpus-based studies (e.g. Plag et al. 2007, 2008) have shown that categorical, rule-based approaches that make use of argument structure (e.g. Giegerich 2004) or semantics (e.g. Fudge 1984) are not able to account satisfactorily for the existing variability. Using data from the corpus studies by Plag and collegues, I argue in this paper that an exemplar-based approach is better-suited to accounting for stress assignment in English noun–noun compounds than a traditional, rule-based paradigm. Specifically, it is shown that two current implementations of exemplar-based algorithms, TiMBL (Daelemans et al. 2007) and AM::Parallel (Skousen & Stanford 2007), clearly outperform comparable rule models in terms of how well they predict stress assignment in the corpora. Furthermore, systematic testing reveals that the reasons for the differences between exemplar and rule models mainly lie in their ability to incorporate detailed, non-abstract information (specifically, constituent family information). The present study therefore adds to the growing evidence in favour of the importance of constituent family information in compounding (e.g. Gagné 2001, Krott, Schreuder & Baayen 2002).

Examining four constructions in three languages (English quantificational nouns, Japanese subordinating conjunctions, Cantonese coverbs, Japanese deverbal postpositions), this paper shows that semantic properties can change faster than syntactic properties in gradual processes of grammaticalization. In each of these cases, the syntactic properties of one category become associated with the semantic properties of a different category when an item undergoes semantic change, leading to the appearance of mixed categorial properties. We propose that this sort of change is best captured using a multi-modular framework (Sadock 1991, Yuasa 2005), which allows changes to affect semantics independently of syntax, and which shows clearly that the relevant items and constructions still conform to the separate structural constraints of syntax and semantics, despite the unusual combination of properties. These findings are important for theories of grammaticalization because they suggest that the cover term ‘decategorialization’ (the loss of grammatical properties associated with the source category) must be understood in terms of at least two separate processes: (1) the effects of semantic change on an item's distribution; and (2) the effects of frequency (Bybee & Hopper 2001) and Pressure for Structure–Concept Iconicity (Newmeyer 1998) on an item's syntactic categorization. Our case studies show that the first kind of decategorialization effects can occur even in the absence of the second kind. Implications of these findings, including possible reasons for both the instability and the long-term retention of mismatch constructions, are also considered.

The positions of subject and object in simple active sentences, and object of a preposition are normally filled by NP's but they can also be filled by PP's. There are reasons for analyzing such PP's as just PP's and not as PP's embedded in NP's. Yet such an analysis appears to be problematic for the analysis of raising and passive sentences assumed within the government-binding (GB) framework, which is formulated with exclusive reference to NP's. Assuming that the basic mechanisms at work in sentences with PP's in typical NP positions are identical to those in sentences with NP's, PP's must sometimes require Case to avoid a violation of the Case Filter. It is possible to incorporate this idea into the GB framework given a particular view of Case and Case assignment, and a revised Case Filter. The analysis receives some further support from a range of related data.

This paper argues that, with syntax defined as progressive projection of semantic representations along the left-to-right dimension provided by the sequence of words (Cann, Kempson & Marten 2005), explanations for local and (multiple) nonlocal scrambling of NPs in Japanese and Korean follow from general principles of tree growth, allowing differences between the languages while nevertheless retaining an integrated account of scrambling itself. This formalism is similar to the parsing mechanism of Miyamoto (2002), but goes further in using this as the base grammar formalism, with all concepts of movement replaced by progressive articulation of structural underspecification and tree growth starting from the left periphery. The account extends the analysis of Japanese scrambling of Cann et al. to encompass multiple long-distance scrambling, capturing both the attendant relative locality restriction on the constituents moved, and interaction of this restriction with scope-construal effects. Scope-construal variability is expressible as interaction between individual lexical specifications for the two languages and general constraints on scope construal; and the relative locality constraint on the construal of the expressions involved in multiple long-distance scrambling is an immediate consequence of the general dynamics of the framework. The resulting account extends Hawkins' (2004) program of defining grammars relative to performance considerations.

Recent research has identified characteristic differences between the phonetic structures of names and of ordinary nouns, with particularly distinctive patterns being exhibited by feminine personal names. No explanation has yet been found. This paper suggests that the solution lies not in the English sound system, as has previously been assumed, but in differences between the linguistic origins of the various types of material.

In this article data are presented from the Amele language and a number of other different Papuan languages to show that in these languages the medial verb form is marked for a binary distinction of realis versus irrealis modality. Furthermore it is demonstrated that in these languages the realis-irrealis distinction interacts with categories of tense and mood marked on the fina! verb in such a way as to divide them into the two basic domains of realis and irrealis modal status. It is also demonstrated for two of the languages (Amele and Nobonob) which have sentence particles with an illocutionary force function that these particles do not interact with the realis-irrealis modal system marked on the medial verb. In the final discussion and conclusion a schema is suggested to account for the facts of future tense, imperative mood and counterfactual mood being categorized as irrealis in these languages. The schema suggests that events can be referred to in the unreal world as either events that are potentially true/false in the real world, or events that used to be true/false in the real world, or events that are not true/false in the real world.

The impetus for this paper is a curious case of gender disagreement in Scandinavian, whereby the rule assigning gender agreement to predicate adjectives is apparently violated. An analysis of the facts is provided, and it is shown that other semantico-syntactic phenomena of Scandinavian can be handled within the same explanatory framework.

Persian complex predicates pose an interesting challenge for theoretical linguistics since they have both word-like and phrase-like properties. For example, they can feed derivational processes, but they are also separable by the future auxiliary or the negation prefix.

Various proposals have been made in the literature to capture the nature of Persian complex predicates, among them analyses that treat them as purely phrasal or purely lexical combinations. Mixed analyses that analyze them as words by default and as phrases in the non-default case have also been suggested.

In this paper, I show that theories that rely exclusively on the classification of patterns in inheritance hierarchies cannot account for the facts in an insightful way unless they are augmented by transformations or some similar device. I then show that a lexical account together with appropriate grammar rules and an argument composition analysis of the future auxiliary has none of the shortcomings that classification-based analyses have and that it can account for both the phrasal and the word-like properties of Persian complex predicates.

Late Middle English is a VO language in which surface OV order can nevertheless be found. This has been taken to show that the language has both OV and VO as possible underlying orders. However, there are various types of data that this hypothesis cannot account for. It is therefore argued in this article that late Middle English only had VO order underlyingly, and that all surface OV orders are derived by means of leftward movement of the object across the verb, in a manner fully compatible with the Minimalist Program. In surface OV, the object moves overtly to the functional projection AgrOP for feature-checking; LF movement is also possible (yielding surface VO), since the two derivations are equally economical. Pronominal objects can move to a higher clitic position. This analysis is shown to account for a wide range of data.

Cappelen & Lepore (2005, 2006a, 2007) note that linguistic communication requires ‘shared content’ and claim that Relevance Theory makes content sharing impossible. This criticism rests upon two important errors. The first is a flawed understanding of Relevance Theory, shown in the application of an omniscient third party perspective to parts of Relevance Theory that depend only upon subjective judgements made by the addressee of an utterance. The second is confusion about different definitions of content. Cappelen & Lepore's evidence actually involves the communication of what they term Speech Act content, which need not be perfectly ‘shared’ according to their own position. Looking beyond this flawed criticism, a general comparison of Relevance Theory with Cappelen & Lepore's semantic minimalism reveals significant parallels, pointing to a notable convergence of two distinct approaches – one cognitive-pragmatic, the other philosophical-semantic – on the rejection of currently dominant assumptions in linguistic semantics. The key remaining difference is Cappelen & Lepore's claim that shared content is propositional. This contradicts other claims made for such content and in any case plays no active role in the explanation of communication. Cappelen & Lepore's position thus poses no threat to Relevance Theory; rather, Relevance Theory can benefit from their philosophical analysis of the state of semantic theory.

This paper discusses cases of syntactic doubling in wh-dependencies attested in dialects of Dutch, where more than one member of the same chain is spelled out. We focus on cases of non-identical doubling, in which the chain links spelled out have different forms. We demonstrate that the order of elements in a chain is fixed: the first (or syntactically higher) one is less specific that the second one. We argue that this generalization follows from partial copying, a process that copies a proper sub-constituent and remerges it higher in the structure. This naturally excludes the ungrammatical orders, as these would involve full copying plus the addition of features, in violation of the inclusiveness condition. The proposal requires pronouns to be spell-outs of phrases, and it is in combination with this hypothesis that the full set of data is accounted for in a uniform way. Advantages over alternative accounts of syntactic doubling are discussed.

It is well known that the quantity indicated by an NP affects clausal aspect if the referent of the NP participates in the event incrementally, i.e. in a part-by-part manner (e.g. She was mowing the lawn). In general, an incremental NP that indicates a closed quantity makes the overall aspect of the sentence telic and thus bounded, whereas one indicating an open quantity results in unbounded aspect (e.g. Water was dripping from the ceiling). In this paper the interplay between quantity and aspect will be called nominal aspect. It is argued that quantity may relate with time in two different ways: first, as overall quantity (which, if incremental, cumulates over time), and second, as transient quantity. The latter term refers to the quantity involved in the situation at a given point in time. It is argued that the interpretation of certain NPs evokes both kinds of quantity; e.g. in This machine pumps the waste water of the factory into the drain the object indicates a quantity that is open in the overall sense (there is no end to the waste water entering the event of pumping) but closed in the transient sense (at any point, all [relevant] waste water gets pumped into the drain). A corresponding distinction is drawn in the domain of verbal aspect, which can also be bounded or unbounded in two different ways. Overall aspect unfolds over time and, if telic, ultimately reaches its endpoint, as in She took the letter to the post office. Transient aspect is the aspectual nature of an event at any given point in time. It is understood as orthogonal to the time axis and gives a cross-section of the ongoing event. In This brush cleans the conveyor belt before it enters the machinery the overall aspect (of the ‘cleaning’) is unbounded, but the transient aspect is bounded, assuming that the brush continuously keeps the conveyor belt in a state of total cleanliness. In this paper, such oppositions are used in explaining the case marking of the Finnish object (partitive vs. ‘total object’ case marking), which reflects both quantificational and aspectual factors. It is argued that the total object can indicate a closed quantity and a bounded aspect not only in the overall sense but also in the transient sense. This distinction is then used to account for many hitherto unexplained uses of the cases.

Two Romanian nominalizations, the infinitive and the supine, are compared in Noun+Object (NO) and Noun+Subject (NS) structures, regarding their ability to yield e-(vent)/r-(esult) readings. The NO structures behave alike and yield e-readings. The two NS structures contrast sharply: the infinitive NS is always an r-nominal, the supine NS may be an e-nominal. This contrast between the infinitive and the supine follows from their aspectual properties. While the supine is [−Telic], and may project either an Object or a Subject in e-nominals, the infinitive is [+Telic], and REQUIRES the projection of the Object. This constraint may follow from the fact that in nominals Aspect and Case are checked in the same projection.

This paper presents new experimental data on Quito Spanish /s/-voicing, which has attracted considerable interest from theoretical phonologists owing to the overapplication of voicing to word-final pre-vocalic /s/. Bermúdez-Otero (2011) singles out Quito /s/-voicing as an important test case for discriminating between two competing theories of phonology–morphosyntax interactions: Output–output correspondence and cyclicity. Overapplication in /s/-voicing cannot be captured using correspondence relationship to a base form, which challenges Output–output correspondence as a theory of opacity. However, the argument only holds insofar as word-final pre-vocalic /s/-voicing is considered phonological, as Output–output correspondence can account for /s/-voicing assuming that it only applies in the phonetics (Colina 2009). We discuss the diverging empirical predictions concerning categoricity and gradience in the surface realisation of voicing processes. We further test these predictions based on acoustic data from seven speakers of Quito Spanish. Evidence from speech rate manipulations shows that some speakers produce more voicing during frication at normal speech rate, compared to fast, maintaining a stable voicing ratio across different speech rates. We argue that for these speakers, /s/-voicing is optional but categorical, and so it ought to be analysed as phonological. This result presents a challenge to the Output–output correspondence approach, but can be accommodated within cyclicity.

Algorithms for the assignment of main word stress in Dutch have never been systematically tested against foot-based segmental processes. The assumption has apparently been that such processes do not exist. In this article, it is suggested that Dutch has at least four segmental rules that make reference to the foot, and that Dutch has a chanted intonation contour whose realization is governed by foot structure. The evidence provided by all these processes largely confirms the more recent proposals for Dutch foot structure, including Kager (1989) and Trommelen & Zonneveld (1989), except where they fail to conform to (1).

In his paper ‘On Horn's dilemma: presupposition and negation’ Burton-Roberts (1989a) presents an ambitious programme, formulated right at the outset. He seeks to establish three points:

(i) Under the ‘standard logical definition of presupposition’ a pre-suppositional semantics is INCOMPATIBLE with a SEMANTICALLY AMBIGUOUS NEGATION operator (SAN), on pain of the theory being rendered ‘empirically empty and theoretically trivial’,.

(ii) From this it follows that the one unambiguous negation is presupposition preserving. Cases that have been identified as presupposition-cancelling negation should be re-analysed as ‘instances of a pragmatic phenomenon’, not unlike what has been proposed in Horn (1985), that is as METALINGUISTIC NEGATION (MN).

(iii) This pragmatic analysis of MN ‘itself implies a presuppositional semantics’, that is to say ‘a presuppositional theory of truth-value gaps’.

This paper investigates the properties of plural agreement that is triggered by collective nouns in British English. Both singular and plural agreement are able to appear with these collective nouns, which are shown to be morphologically singular but semantically plural. Plural agreement, however, is systematically more restricted than singular agreement, appearing in a subset of the environments where singular agreement is allowed. Restrictions on plural come from the nature of agreement; semantic agreement features can only enter into agreement when the controller of agreement c-commands the target of agreement, whereas morphologically motivated agreement is not subject to the same structural restriction. This asymmetry between the two types of agreement is shown to arise from the proposal that Agree (Chomsky 2000, 2001) is distributed over the syntactic and post-syntactic components (Arregi & Nevins 2012).

The key assumption in the standard generative approach to historical linguistics (King, 1969) is that each sound change is incorporated directly into the native speaker's grammar as the final phonological rule, moving up gradually into the grammar as further changes are implemented. Restructuring of underlying representations by later generations during acquisition is theoretically permitted, but infrequently invoked, with the result that the historical phonology of a language will be almost directly mirrored in the order of its phonological rules. The only extractable generalizations are then that the ‘highest’ rules will correspond to the oldest changes, and that a sound change and the rule into which it is converted will tend to be identical or at least show a high degree of similarity in formulation. This approach casts no light at all on the problem of the implementation of sound change.

In his (1970) article ‘Surface structure constraints in syntax’ Perimutter has argued that the restrictions on the combination of the weak atonic forms of the object pronouns in Spanish and restrictions on their order cannot be accounted for naturally either in the deep structure, by constraints on the phrase structure rules, or by constraints on the rules which introduce these pronouns.1 He concludes that a transformational generative grammar must be supplemented by positive output constraints which act as a filter accepting only sequences which match the filter and rejecting the rest as ungrammatical.