Construction Grammar Week 10: Phrasal Constructions

This thesis explores the difference between separable and non-separable transitive English phrasal verbs, focusing on finding a reason for the non-separable verbs’ lack of compatibility with the word order alternation which is present with the separable phrasal verbs. The analysis is formed from a synthesis of ideas based on the work of Bolinger (1971) and Gorlach (2004). A simplified version of Cognitive Construction Grammar is used to analyse and categorize the phrasal verb constructions. The results indicate that separable and non-separable transitive English phrasal verbs are similar but different constructions with specific syntactic reasons for the incompatibility of the word order alternation with the non-separable verbs.

Language, 2007

Start­ing in the nineties more and more lin­guis­tic ar­ti­cles were pub­lished in the frame­work of Con­struc­tion Gram­mar. Al­though Kay and Fill­more (1999, p. 19) made it clear that Con­struc­tions are not nec­es­sar­i­ly phrasal, most of the au­thors sug­gest phrasal Con­struc­tions. This is es­pe­cial­ly ap­par­ent in Con­struc­tion Gram­mar-​in­spired work in the frame­work of HPSG. In this paper, I show that the dif­fer­ence be­tween phrasal ap­proach­es and lex­i­cal ap­proach­es is not as big as it is some­times claimed, but that the de­ci­sion for one of the ap­proach­es nev­er­the­less may have se­ri­ous con­se­quences. The dis­cus­sion fo­cus­es on re­sul­ta­tive con­struc­tions, a phe­nomenon for which both phrasal and lex­i­cal anal­y­ses were sug­gest­ed. I show that an enor­mous amount of dif­fer­ent Con­struc­tions is need­ed to ac­count for all pat­terns that may arise be­cause of re­order­ing of con­stituents or re­al­iza­tion of the re­sul­ta­tive con­struc­tion in con­nec­tion with va­lence chang­ing pro­cess­es. It will be shown that ad­juncts, pred­i­cate com­plex­es, and deriva­tion­al mor­phol­o­gy pose con­sid­er­able prob­lems for the phrasal ap­proach, while they are un­prob­lem­at­ic for lex­i­cal rule-​based ap­proach­es.

Handbücher zur Sprach- und Kommunikationswissenschaft / Handbooks of Linguistics and Communication Science (HSK) 42/1, 2015

Table of contents 1. Introduction 2. Examples of constructions 3. Constructions and syntactic theory 4. Conclusion 5. References (selected)

The aim of the present chapter is to focus on the most complex and schematic end of the constructicon cline. In particular, I will look at Abstract Clausal (declarative, interrogative, imperative and relative clauses; cf. section 1) and various Filler-Gap constructions (such as topicalised clauses, e.g. [My bagels]_i, she loves_i, or wh-interrogatives, e.g. [What (books)]_i do they love_i?, in which a constituent (here [My bagels]_i and [What (books)]_i) appears in a nonargument position; cf. section 3). Moreover, I will show how in English the Filler-Gap construction interacts with preposition stranding (Who did she talk to?) and pied-piping (To whom did she talk?; cf. section 3.1), and how the complexities of the English relative clause constructions can be accounted for (section 3.2). On top of that, I will also illustrate how intonation can be an integral form feature of constructions (e.g. the English Comparative Correlative construction The X-er the Y-er; cf. section 3.3). Besides, contrary to the prejudice that Construction Grammar only focuses on peripheral and idiosyncratic exceptions, throughout the chapter, I will not only show that constructionist approaches can describe and model the most abstract and schematic of syntactic structures, but that such analyses also further our understanding of the phenomena in question. Please find a pre-publication draft of the chapter below.

Journal of Linguistics, 2021

2015

Review of Martin Hilpert (2014) Construction Grammar and its Application to English published in 2015 in Poznań Studies in Contemporary Linguistics 51(4): 605-613.

Constructional Approaches to Language, 2004

Every construction is associated with more or less detailed information about its phonological, morphological, syntactic, semantic, pragmatic, discourse, and prosodic characteristics. Since such characterizations may swell into fairly large and elaborate collections of symbols when represented formally, Construction Grammar uses a box notation as a convenient way of organizing all the information needed to give an adequate account of linguistic structure. The box diagrams have become the most visible and readily recognizable trademark of Construction Grammar representations. Constructions have always played an important role in grammars and linguistics; traditionally, we talk about sentence types, phrases, formulas, and even idioms. In Construction Grammar, the notion of 'knowing a language' means knowing its constructions; the active, the passive, the reflexive, the existential sentence types can all be seen as constructions, and so can the preposition phrase, or the verb phrase. In fact, in the view of Construction Grammar, language is the inventory of its constructions. 2.2 The Case Grammar connection As noted in the introductory chapter, Construction Grammar evolved out of Case Grammar (Fillmore, 1968, 1977; Dirven & Radden, 1987) and the early versions of Frame Semantics (Fillmore, 1982, 1984). Case Grammar was one of the first approaches that set out to search for a semantically defined 'deep structure' and its manifestations in linguistic expressions. 1 Thus, the primary reason for saying that John Smith has a different semantic role in (1a), below, than England in (1b) is not the inherent and intuitive difference in meaning between a person and a country, but the fact that the two display different syntactic behavior; in nominalizations, for example, one tends to take the s-genitive, and the other the preposition in, as shown in (2a) and (2b), respectively. (1) a. John Smith remembers nothing of years gone by. b. England remembers nothing of years gone by. (2) a. John Smith's memory of years gone by is non-existent. b. The memory of years gone by is non-existent in England. When comparing the noun God used in (3) below to either John Smith or England, we notice that it patterns syntactically after John Smith (shown in 4a), or at least more so than after England (shown in 4b), even though intuitively, based on its referential properties, God might seem distinct from either of the other two nominals. (3) God remembers nothing of years gone by. (4) a. God's memory of years gone by is non-existent. b. ? The memory of years gone by is non-existent in God. On the basis of these facts we might want to assign the same semantic role, say, 'agent', to John Smith and God, but a different role, say, 'location', to 4 England (cf. Fillmore's 1971 arguments against the need for a semantic role 'force'). Similarly, we can deduce that the word children in (5a) is, at least in principle, semantically ambiguous, since in a passive sentence with an oblique adverbial, we have to choose between using the preposition by (which indicates that children functions as agent, as in 5b) or with (indicating that children is an 'instrument', as in 5c); compare to (5d) which contains both agent and instrument roles. (5) a. Children filled the bewitched house. b. The bewitched house was filled by children. c. The bewitched house was filled with children. d. The bewitched house was filled with children by the unscrupulous witch. Fillmore (1968) explicates the regularities in mapping semantic roles onto different grammatical functions in sentences. Thus, in English, if there is an agent in an active sentence, that agent is realized as the subject; if there is no agent, but an instrument, the instrument is realized as subject; and if there is no agent nor instrument, but something that is affected by an activity, a 'patient', then the patient is realized as subject. This is illustrated in (6). (6) a. The Chancellor closed the university with a dull speech. b. A dull speech closed the university. c. The university closed. The semantic role patterning is still at the core of Construction Grammar. In early studies in Frame Semantics, Fillmore developed the notion of roles further, suggesting that grammar can be seen as a network of associations between syntactic roles (more generally known as grammatical functions), textual roles (accounting for information structure), and verbspecific situational roles (such as 'buyer' and 'seller' in a commercial transaction). These relationships will be addressed in section 6. 3. Arguments for Construction Grammar Although the physical realization of language (what we see as form and hear as sound) is what comes closest to being observable and thus empirically based, there are very few, if any, patterns in English that can be said to be purely syntactic, in the sense that their meaning or function play no role in determining well-formedness. The closest we come to a purely syntactic pattern may be what is known as the Subject-Predicate construction, since almost anything can be the subject in English. Most often, however, a construction has among its defining properties specific semantic and pragmatic features. It is not uncommon that even when the structure of two phrases seems to be exactly the same, as in the expressions Thank you and See you, the two expressions may 6 The relation between 'productive rules' and 'idioms' must be seen as a cline from relatively productive to relatively frozen. There is no sense in treating the constructions of a language as belonging to qualitatively different categories on the basis of their degree of productivity. True, there are idioms that benefit little from being integrated into the productive parts of grammar; for instance, by and large or trip the light fantastic are clearly at the frozen, formulaic end of the scale. But even so, they are not completely without tractable structure. In expressions such as What's Bill doing inspecting the car? or What's it doing snowing in August?, as discussed in Kay & Fillmore (1999), or the greener the better (Fillmore, 1989), it is not at all clear whether it is more appropriate to treat these as idioms, or as productive kinds of structures. Construction Grammar does not have to make that choice. Another area that illustrates a gradient scale between the formulaic and the productive is that of numbers. Although it may seem that numbers are to a certain extent 'peripheral', they are clearly part of our language and they commonly make up systems that are subject to general grammatical constraints and thus form an integral part of grammar. This is readily apparent in a language like Finnish, where numbers partake in concord relations. In order to say 'in 35 rooms', Finnish speakers do not say '35 room-in', as in (7c), with the number specification in an unmarked, default case, but, minimally, '35-in room-in', as in (7b), and preferably in the form kolmessakymmenessäviidessä huoneessa '3-in 10-in 5-in room-in', as shown in (7a). 2 (7) a. kolme-ssa-kymmene-ssä-viide-ssä huonee-ssa three-IN-ten-IN-five-IN room-IN b. kolmekymmentäviidessä huoneessa c. *kolmekymmentäviisi huoneessa Evidently, numbers are like other nominals in Finnish in that they are assigned case suffixes. But numbers are not entirely like any other nominals, either; they have their own characteristics that need to be captured in a full account of language. For instance, when accounting explicitly for how numbers are made up morphologically in English, we need to invoke a set of principles that are not frequently referenced elsewhere in the English grammar. In particular, speakers use addition to form sequences like seventeen: 7 + 10 ('seven plus te(e)n') or twenty-three: 20 + 3 ('twenty plus three'), but multiplication is used in forming seventy: 7 x 10 ('seven times ten [=ty]'). We can argue that in Finnish, subtraction is also at play, e.g. kahdeksan 'two away from ten', i.e., 10-2 = 'eight'. And neljätoista 'four of the second' displays a complex structure involving both multiplication and addition to designate 'fourteen'. The discussion of numbers also points to the inherent similarity between word-length and sentence-length constructions: Construction Grammar does not have to make an a priori choice of whether to consider a piece of linguistic material (in this case, any number) a word, a phrase, or a

2007

In this paper I suggest an interface level of semantic representations, that on the one hand corresponds to morpho-syntactic entities such as phrase structure rules, function words and inflections, and that on the other hand can be mapped to lexical semantic representations that one ultimately needs in order to give good predictions about argument frames of lexical items. This interface level consists of basic constructions that can be decomposed into five sub-constructions (arg1-role, arg2-role ... arg5-role). I argue in favour of phrasal constructions in order to account for altering argument frames and maybe also coercion without having to use lexical rules or multiple lexical entries.

In Construction Grammar, grammatical patterns are conventional pairings of form and meaning that are analogous to words. This article contrasts Construction Grammar with competing syntactic theories that are based on universal constraints and the projection properties of words. It reviews arguments for construction-based syntax derived from the following linguistic phenomena: semantic and syntactic variability of verbs, coercion, idiomatic patterns and ‘family resemblances’ among idioms, paradigm-based constraints on form and meaning, exceptions to cross-constructional generalizations, and the inadequacy of derivational rules. Verbal and nominal syntax are used to exemplify the formal mechanism that combines constructions and words, unification grammar. A concluding section outlines connections between Construction Grammar and use-based models of grammar, acquisition and sentence processing.

To appear in: Barbara Dancygier, ed. The Cambridge Handbook of Cognitive Linguistics. Cambridge: Cambridge University Press., 2017

To appear in: Barbara Dancygier, ed. The Cambridge Handbook of Cognitive Linguistics. Cambridge: Cambridge University Press. The previous chapter gave an overview of the renaissance of constructions in grammatical theory and the rise of Construction Grammar approaches. Yet, while all constructionist approaches share many important tenets concerning the nature of human language, the various individual approaches nevertheless differ from each other in non-trivial ways. In this chapter, I will first provide the common theoretical assumptions shared by all constructionist approaches. After that, I will outline the major differences between non-usage-based (such as Berkley Construction Grammar and Sign-Based Construction Grammar) and usage-based approaches (Cognitive Construction Grammar, Embodied Construction Grammar, Fluid Construction Grammar and Radical Construction Grammar). Moreover, I will discuss the controversial issue of what counts as a construction (from Kay's conservative competence-based notion to the usage-based interpretation of constructions as exemplar-based clouds) and the ontological status of meaningless constructions. In addition to that, I will also touch upon the nature of the structured inventory of constructions, the constructicon, and explore the advantage and limits of constructional inheritance in taxonomic networks. Finally, the chapter will also address the question as to how the meaning pole of constructions is analysed in the various approaches (which ranges from semantic paraphrases (Cognitive Construction Grammar) over first-order predicate logic (Fluid Construction Grammar) to Frame-based approaches (Sign-based Construction Grammar)).