Belgian Journal of Linguistics - Volume 31, Issue 1, 2017

Over the last 20 or so years, research on syntactic alternations has made great strides in both theoretical and methodological ways. On the theoretical side, much of the research on syntactic alternations was restricted to generative linguistics debating how near synonymous constructions differed slightly in meaning and/or how one (and which one) was derived from the other (transformationally). On the methodological side, much research consisted of monofactorial studies based on relatively simple text counts. By now, however, syntactic alternation research has become much more functional (in a broad sense of the term) and much more methodologically sophisticated: Much work is now motivated/interpreted psycholinguistically or in a broadly usage-based/cognitive linguistic framework and much work has now adopted a regression-based analytical strategy. These attractive developments notwithstanding, much remains to be done and, in this paper, I sketch some recent developments in (largely) separate alternation studies that I would like the field to adopt more broadly. These developments can be heuristically grouped into ones that have to do with (i) the statistical analysis of corpus-based and experimental alternation data, (ii) new predictors that explain typically unexplored aspects of variability in alternations.

In this paper, I examine whether the variation patterns of haber pluralization (e.g., hubo/hubieron fiestas ‘there was/were parties’) in Peninsular Spanish corroborate the hypothesis elaborated in earlier work that the phenomenon constitutes a competition between two variants of the presentational construction with haber that is constrained by domain-general cognitive constraints on spreading activation. In addition, this paper examines whether haber pluralization is incrementing in frequency in particular Peninsular regions and whether or not the phenomenon is spreading geographically. To meet these objectives, I analyze a dataset of more than 7,500 cases of haber + plural NP, which were culled from two publicly available data sources: the Corpus Oral y Sonoro del Español Rural (which represents only rural speakers born before the 1940s; Fernández-Ordóñez 2005- ) and Twitter (which represents mainly young and middle-aged speakers). The results of a mixed-effects logistic regression analysis that tests the effects of tense, the absence/presence of negation, typical action-chain position of the noun, the regional origin of the examples, and the data sources support the competition hypothesis. This model also supports that pluralized haber is spreading westward from its epicenters (Valencia, Barcelona, and Murcia), while also incrementing in frequency in northern, eastern and southern Spain. However, its frequency appears to be declining in central Spain. A geographically more detailed, but similar picture is obtained with three generalized additive mixed models that test the effects of geography on the total dataset as well as on each of the two subcorpora.

Our contribution is dedicated to the empirical testing of alleged regional variants of object clitic clusters in modern French in France, Belgium and Switzerland. We provide some intriguing new insights into the regional distribution of non-standard variants and discuss one hypothesis on their nature and two hypotheses to explain their coming into being: language-contact (with Francoprovençal, Occitan and Oïl dialects, H1) and/or analogical leveling (H2), on the one hand, and their postsyntactic, rather than syntactic, nature, on the other (H3). Our main results reveal that the three non-standard variants where order in object clitic clusters is concerned are not regionally well-distributed, i.e. the observed distribution does not correspond to any cohesive area. In contrast, only one variant where the selection of the form (me vs. moi) is at issue seems to be regionally confined: it is found in French-speaking Switzerland, in Gascony, plus some rare attestations of it in the North of France. All in all, variation in object clitic clusters indicates a genuinely new geographical articulation of regional French that does not coincide with traditional dialectal areas.

The causative-inchoative alternation has been a subject of much debate. It might also be a case where variation patterns that escape existing typological descriptions provide a new perspective on the problem. We analyze the variability and systematicity of alternative argument structure realizations, together with corresponding aspectual/event properties, by considering three different ways in which change-of-state verbs can be semantically and syntactically construed in Romance. Under the general assumption that the syntactic projection of arguments correlates non-trivially with event structure, we apply a novel theoretical approach to the semantics and syntax of the causative-inchoative alternation. We argue that different verbal heads can be independently combined to yield contrasting verbal configurations, with corresponding event/argument structure properties quite freely. Alongside standard cases such as causative and inchoative frames, we discuss what we call ‘stative-causative constructions’ [SCC], where the initiator appears as the sole argument. The general properties of this additional (third) variant suggest the availability of a null causative (external-argument-selecting) v0 producing original monoargumental structures with corresponding (simpler) event structure. These little-known Spanish data challenge current argument structure theories assuming that the causative v0 necessarily implicates the eventive (BECOME) component, or that the latter figures in the verb’s permanent lexical entry. SCCs provide empirical evidence suggesting that what is commonly described as a basic unaccusative/transitive verb may have unergative uses.

This paper investigates agreement mismatches in Dutch relatives. While the norm is that singular neuter nouns occur with the relative pronoun dat ‘that’, it is by now quite common to find neuter nouns combining with the relative pronoun die. A large Twitter corpus is used to study which linguistic variables make die ‘that’ in this context more likely. Lack of agreement between neuter noun and relative pronoun is very frequent in this corpus (37.5% of the cases, 46.8% if the preceding determiner is indefinite). Non-agreement is most common for nouns that are high in the animacy ranking, but it also occurs with other semantic classes, and there is quite a bit of lexical variation. Young, female users have a stronger tendency to use non-agreeing relative pronouns. Contrary to what previous work suggests, we do not find that users with a Moroccan or Turkish background have a stronger tendency towards non-agreement. A comparison of tweets with agreeing and non-agreeing pronouns and a comparison of the Twitter corpus with web data both suggest that non-agreement is characteristic of informal language use.

Several syntactic properties of verbal heads are accounted for through their semantic properties. Verbal features such as agentivity, volitionality, stativity etc. have been proven a useful tool for predicting several aspects of their syntactic behavior such as passivization, auxiliary selection etc. In the context of the empirical turn in current linguistics, the assumption of discrete features is questioned by studies based on corpora or speakers’ intuitions showing that the diagnostics of semantic features involve gradience. These findings are challenging for grammatical theory: are we justified to assume the existence of discrete verb classes or do the established properties indicate scalar dimensions of meaning? Based on two empirical studies – an acceptability study and a corpus study – the present article examines the role of agentivity in distinguishing verb classes and in predicting the syntactic behavior of verbs in German. Acceptability data show that the diagnostics of agentivity involve gradience, which cannot be reduced to random sources of variation. However, a comparison of scalar vs. categorical models of agentivity based on these diagnostics reveals that the syntactic variation in word order found in written corpus data is best accounted for through a model that assumes a binary division into a ±agentive and a non-agentive verb class.

This paper investigates the conditions that govern the choice between the German neuter singular relative pronouns das ‘that’ and was ‘what’. We show that das requires a lexical head noun, while in all other cases was is usually the preferred option; therefore, the distribution of das and was is most successfully captured by an approach that does not treat was as an exception but analyzes it as the elsewhere case that applies when the relativizer fails to pick up a lexical gender feature from the head noun. We furthermore show how the non-uniform behavior of different types of nominalized adjectives (positives allow both options, while superlatives trigger was) can be attributed to semantic differences rooted in syntactic structure. In particular, we argue that superlatives select was due to the presence of a silent counterpart of the quantifier alles ‘all’ that is part of the superlative structure.

This paper presents a method for quantitative and qualitative analyses of the causative alternation in English, where verbs may alternate between a transitive (causative) construction (Santa crinkled his eyes) and an intransitive (non-causative) construction (His eyes crinkled). 1 The aim of this paper is to present a method designed to measure the alternation strength of causative verbs, i.e. the extent to which they alternate between the two constructions. One of the central elements this paper investigates is the Theme, i.e. the participant that is in subject position in the intransitive construction and object position in the transitive construction. A distinctive collostructional analysis ( Gries and Stefanowitsch 2004 ) shows that certain verbs are significantly attracted to one of either two constructions while others are equivalently distributed in the two constructions. However, after careful analysis it appears that very few Themes actually overlap between the two constructions ( Lemmens forthcoming ) which indicates that each construction seems to be rather restrictive regarding which Themes they recruit. The low degree of alternation of the Themes leads us to ask ourselves the extent to which the alternation is part of a speaker’s knowledge of their language.

This paper addresses the early variation in what has been called the [prep_que] variable in Spanish nominal complement clauses, i.e. the alternation between de que and que in examples such as en señal (de) que lo estimo, Zulema, este anillo ofrezco (CORDE) ‘as a sign that I appreciate You, Zulema, I offer this ring’. By applying several subsequent quantitative analyses on corpus instances of the sequences N de que and N que, the locus of variation is restricted to such an extent that the variation can largely be accounted for. A collostructional analysis identifies 31 central nouns of the N de que complement clause construction. A diachronic cluster analysis delimits the temporal dimension of the variation to the 16th and 17th centuries. A distinctive collexeme analysis identifies nine nouns which are used in both constructional formats to a comparable degree: causa ‘cause’, duda ‘doubt’, esperanza ‘hope’, fe ‘faith’, opinión ‘opinion’, recelo ‘fear’, señal ‘sign(al)’, sospecha ‘suspicion’, and temor ‘fear’. Detailed contextual analysis of the use of these nine nouns by means of a mixed-effects logistic regression reveals that the use of the nouns with a determiner is correlated with the de que variant, and the use of the nouns as part of complex predicates, as in tener sospecha ‘have suspicion’, is associated with the que variant of the complement clause.

The present work studies the grammatical properties of Spanish time constructions involving hacer ‘to make’ from an empirical, quantitative perspective with insights from both competence and language use. The results show a big difference between the properties of the so-called clausal and adverbial constructions, something unnoticed until now. Whereas hacer, in the former, has properties not significantly different from other verbs, in the latter, it shows a strong erosion of its verbal and clausal properties like TAM morphology, negation, time adjunction or word order freedom. This contradicts earlier contributions to the topic and argues against those formal proposals positing a synchronic derivational relation between the two constructions.

This study deals with the complexity of left-dislocated [LDed] noun phrases in the Modern English period (1500–1914). The purpose is twofold: to estimate the effects of a number of theoretically relevant predictors on complexity, operationalized as word-length, and to explore whether shorter LDed NPs, which are characteristic of contemporary spoken English, symbolize the claimed ‘orality’ of earlier speech-related texts. Multiple linear regression is employed in order to assess the correlation between complexity and factors relating to both the context (e.g. genre) and the inner configuration (e.g. augmentation of the LDed constituent) of the tokens extracted from the Penn-Helsinki suite of corpora. The results confirm that shorter LDed items more closely match the characterization previously offered for contemporary spoken Left Dislocation. The results also provide preliminary insight into the accommodation strategies that might facilitate the resumption of heavier LDed NPs. As regards genre, speech-purposed and mixed texts feature a higher relative rate of shorter LDed elements. Thus, if shorter LDed NPs are more closely connected with the spoken variant, it is in speech-purposed and mixed genres where they might be claimed to constitute a trait of orality that is deliberately employed by writers in order to reproduce conversation. Surprisingly, speech-like genres do not include LDed items as often, and these are just as complex as those attested in writing-related texts.