This book originated as a series of lectures that I gave in Beijing as part of the 15th China International Forum on Cognitive Linguistics in May 2016. It offers an account of events and event structure within a theory of language, Word Grammar, which assumes that language is a cognitive network with no embedding where the language network is just part of the larger cognitive network. Word Grammar’s language network is a symbolic network which is designed to be as simple as possible: formally, it is just a series of arcs and nodes, classified by default inheritance. The problems that I am concerned with are to do with events and event structures. In a cognitive approach, events are just verb meanings, but an event is not (only) the sense of an individual verb. Events are individually identifiable, with times, places and participants, and these pieces of information have to be built into the linguistic representations. I look at how this information is built up, and its consequences for the representation of events, and the relationship of event semantics to sentences and utterances.
There are several issues in the analysis of events. For example, what is an event? Verb meanings make up a heterogeneous class. Intuitively, we can see that defining the meaning of the word Event as the class of verbs’ meanings makes sense: Explode, Kiss, Run, and Walk all have dynamic states of affairs as their meanings.1 They also have pretty simple meanings. But what do we make of more complex states of affairs? Verb meanings can involve more than one event, so how should we analyse these complex event structures? Does Kill mean something like ‘cause to die’? How can the meanings of stative verbs such as Love be “events”? Other stative verbs include modals like May and copular verbs such as Seem—how do they work? How many participants can an event have? For example, there are two in transitive verbs, but what about ditransitives, or verbs with even more participants such as Buy in he bought it from her for £50, which appears to involve four participants. What are the limits on how events are structured? Are all results entailed? Are there restrictions on how events can combine within verb meanings?2
In the course of these lectures, I tackle these and other related questions while building up a model which allows us to bring the questions to the fore and explore them in a precise way. Word Grammar views language as part of a larger cognitive network—so I explore the consequences of the network architecture and allow it to guide the analyses of event behaviour and event structure. There is some comparison with other approaches, most usually Cognitive Construction Grammar, which has a number of similarities to Word Grammar, except that it permits the recursive embedding of constructions within constructions.
The main argument in the book is that we should take the notion of language as a cognitive network seriously. I think that this obliges us to reject phrase structure, and other devices which require there to be embedding of one structure within another. I argue that knowledge of language is just part of a larger cognitive network, and that the language network is supported by social understanding, and our theory of mind. Even complex event structures can interact with the mental representations of utterances and utterance context, which suggests that language should be viewed as integrated with social cognition and is another argument for the network approach.
With the exception of Lecture 7, each lecture addresses a different theme, rather than being organized around particular data sets, which means that the same data points can recur. For example, ditransitives are discussed in Lecture 7, and then again in Lectures 8 and 9, although with a different take. However, when the data sets recur, it is in order to make a different point: in one case I am interested in the syntax of ditransitives; in another, I look at the semantic relationships between ditransitives and the related prepositional uses of the same verbs.
Lecture Titles and Summaries
-
A network model of language structure
It is widely accepted in cognitive theory that language is a conceptual network. Often, this is understood as a network of constructions. In this lecture, I argue that it makes sense to take this position to its logical conclusion, and to treat language as a symbolic network with no part-whole relations at all: language consists of nodes, and relations, and nothing much else.
-
Parts, wholes, and networks; idioms; semantics—syntax—morphology
This lecture explores the consequences of adopting the position in Lecture 1 for theories of syntax, semantics and morphology, and their interactions. We look at the treatment of idioms as a case study and the lecture finishes with a discussion of WG lexical entries.
-
Evidence for structure in verb meaning
One of the tasks of work on event structure is to establish the nature of the relationship between syntax and semantics. But there is a task that has to happen first: we need to explain why we assume that there is an event structure. What is the evidence that events are complex and that the complexity is structured? What facts can we draw on? Are we obliged to assume that there is structure, or from a theoretical point of view, is it possible to do without it? I argue that there is, indeed an event structure, but that within Word Grammar’s network model, that structure does not involve recursive embedding of one concept within another.
-
Polysemy and semantic structure
One thing a lexical semantic theory has to do is to model polysemy. Some polysemy is relevant to event structure: the polysemy to do with variation in argument realization. But there are other dimensions of polysemy, which don’t affect argument realization: are they relevant to the theory of events? Are there different kinds of polysemy? How should we model polysemy?
-
Events and thematic roles
Do different event types define different thematic role types, or is it the other way around? What diagnostics could we deploy to find out? I argue that we can distinguish fruitfully between Talmy’s “force-dynamic” relations, and the other kinds of semantic role which are usually defined in terms of how language structures space.
-
Classes of event
What categories of event are there? In this lecture we establish a type hierarchy of events, and ask how we might model event types cross-linguistically, looking at Talmy’s claims about verb-framed and satellite-framed languages, and Levin and Rappaport Hovav’s claims about manner/result complementarity.
-
Building structure: ditransitives and verbs of buying and selling
Ditransitive verbs are a complex type: they involve multiple events, and an additional participant—albeit one that has a stable interpretation in the verb’s meaning. Verbs of buying and selling build on the meanings of ditransitives, and here we look at how they require a fine-grained interpretation. This brings us back to the material of the second lecture, in that we explore how the network approach takes us to different conclusions from those you would arrive at by following a frame-semantic approach.
-
Events and Aktionsart—modelling the structure
This lecture builds on the earlier lectures, and explores the issue of how the different Aktionsarten—states, simple dynamic events, semelfactives, achievements and accomplishments—are built up from the tools we have introduced so far. We look at complex events, the possible roles of different thematic role types, and the representation of polysemy.
-
Transitivity alternations and argument linking
One of the key roles for events in linguistic theorizing is in establishing a theory of argument linking or argument realization. In this lecture we explore how a declarative theory of argument linking can be described in terms of the event structures we have established in the earlier lectures, but without recourse to phrasal constructions.
-
Situating meaning in the utterance: modality
As an envoi, this lecture discusses how the intramental-network approach allows us to model how context interacts with the linguistic material. We explore how the subjectivity of modal meaning interacts with the speaker and the hearer, and explore whether this account is compatible with the earlier theory of event structure and argument realization that we have produced. We also look at other evidence for the communicative and social embedding of linguistic information.
The lectures I gave in China were recorded and transcribed. What follows is edited and corrected versions of the transcriptions. As well as removing the oral features of the lectures—the ums and the ers—I have also taken out the thanks to my hosts at the beginning of the lecture, stories and jokes that do not work on the written page even if they might have helped keep the audience engaged, and I have improved the exposition where I thought I was not clear enough or got things wrong. However, the organisation and structure of each lecture follows the spoken version.
I owe a debt of gratitude to many people. Professor Thomas Li Fuyin, invited me to give these lectures, and provided exemplary hospitality during my trip to China. His team of postgraduate students, particularly Jia Hongxia, Li Jinmei, and Du Jing, were very helpful and well organized. I am grateful to the authorities of Professor Li’s university, Beihang University—as well as the other universities in Beijing which hosted these lectures, Beijing Forestry University, Beijing Language and Culture University, Peking University, Tsinghua University, and the University of the Chinese Academy of Sciences—for their practical support and hospitality. The lectures attracted a stimulating audience, and I would like to thank the audience members for their time and for their thoughtful questions. Maarten Frieswijk and Elisa Perotti at Brill have been very patient, and helpful, editors. And I am very grateful to Brill’s indefatigable Production Editor, Fem Eggers. I also owe a debt of gratitude to Dick Hudson who read through the slides and discussed them with me before I travelled to China and who also make helpful comments on the final manuscript submission. The Moray Endowment Fund of the University of Edinburgh gave me a small grant which paid for Yueh-Hsin Kuo to work as a research assistant on this project: I am grateful both to the fund for their support, and to Yueh-Hsin for his work. Finally, I would like to thank my wife, Caroline Lewis (who sadly was not able to travel to Beijing with me as we’d planned), for her support and love.
A note on representational conventions. I use Small Caps for lexical entries or lexemes. I use italics to show examples of utterances. Single ‘quotation marks’ identify the conceptual structures associated with meaning. Double “quotation marks” are reserved for scare quotes and for direct quotations. The labels for syntactic relations such as Subject and semantic relations such as Agent, have a capital letter.
Another complexity is that the word Event is polysemous: it can mean what are sometimes called “eventualities,” including both stative and dynamic verb meanings, and it is also sometimes restricted to the dynamic cases only. I use it in both ways in these lectures, although I have tried to make it clear which interpretation I intend.