This quote from Patrick Hanks reflects very closely the spirit of this volume that tackles the relation between word meaning and human linguistic creativity. We are interested in pursuing the view that words are rich repositories of semantic information that people use to talk about the world in a potentially infinite number of ways. Our goal is to tackle the essence of words insofar as they provide a window onto human cognition and the compositional nature of thought. It is thus imperative that a theory of language addresses how lexical items contribute to the peculiar human ability that goes under the label of “linguistic creativity.”

It is undeniable that words have “meanings” that go above and beyond the scope of linguistic research: They often carry the weight of a person's own experience. We are not aiming at exploring the unexplorable, but in proving that, given an appropriate set of methodological tools, a formal modeling of word meaning and linguistic creativity can be achieved. Linguistic creativity is a “generative” ability to extend the expressive possibilities of language in a potentially infinite number of ways. From the perspective of the lexicon (i.e., word meaning), it is the ability to give new meanings to words beyond their literal use.

As such, the overall task follows the strategy of contemporary generative syntax, which has achieved a basic understanding of how speakers produce and understand novel utterances and has brought simplicity to the great complexity underlying the syntactic structure of sentences.

The contributions in this section are centered around a set of common themes:

1. developing a theoretical vocabulary sufficiently rich to understand how word meanings compose;

2. developing and motivating frameworks for lexical semantics with explanatory force;

3. analyzing cross-linguistic data for achieving linguistically independent models.

Although each contribution approaches the problems from different angles and different data sets, they all highlight the richness of the information carried by words in context. The real challenge, as it emerges from the papers, is whether it is possible to establish a clear boundary between people's words and people's worlds or experiences. The goal is to understand whether there is a level of representation that is independent of specific contextual variations, while accounting for the novel use of words in different contexts.

The authors reach different conclusions: some reject structured representations of lexical information, others show that it is precisely in the structuring of the lexicon that we can achieve an understanding of how meaning changes in context.

The first paper by James Pustejovsky presents recent developments in GL, focusing on the role of qualia structure as a syntax for creating new concepts. The paper addresses fundamental questions on the well-formedness of concepts, the combinatorial possibilities within a generative mental lexicon, and how these principles are motivated by linguistic evidence.

The contribution by Jacques Jayez focuses on the meaning variations of the French verbs “suggerer” (suggest) and “attendre” (wait).

Abstract

Introduction

Recent work about the relation between lexical items, context, and interpretation has highlighted two notions of underspecification (see van Deemter and Peters, 1996 for various points of view). In some approaches, underspecification amounts to code ambiguities in an efficient way, to avoid carrying a set of alternatives during the interpretation process (Reyle, 1995). In the domain of the lexicon, underspecification sometimes takes the form of information enrichment.

Introduction

Metaphor, and the distinction between the figurative and the literal uses of language, have puzzled philosophers and linguists at least since Aristotle. The puzzle can be stated in the following, rough, form: How can words in certain configurations mean something different from what they mean in their literal use, prescribed by the rules of the language, and at the same time convey significant insights into what we, in a given context, take as parts of reality? In order to appreciate the force of the question we must separate the metaphorical meanings from the new literal meanings that an individual, or a group, might introduce into a language, such as parenting, or critiquing. Such innovations are extensions of literal language, not metaphors. Metaphors rest on rules of language, but also violate them. They do not describe reality directly. Thus, the true/false dichotomy does not apply to them without qualifications. An adequate theory of metaphor should explain this unique position of metaphoric meaning. To expand on this a little, this essay proposes that a theory of metaphoric meaning should account for the following list of facts or intuitions.

1. (i) Metaphors give us new meanings and a deepened understanding of the objects of our descriptions or reasoning.

2. (ii) Metaphors can have aesthetic value.

3. (iii) At some stage, a subjective element enters into the interpretation of metaphors. This element is creative insofar as it goes beyond what is given by the rules of language but it presupposes and rests on such rules.

4. […]

Abstract

Language and Category Formation

Since the early days of artificial intelligence, researchers have struggled to find a satisfactory definition for category or concept, one which both meets formal demands on soundness and completeness, and practical demands on relevance to real-world tasks of classification. One goal is usually sacrificed in the hope of achieving the other, where the results are muddled with good intentions but poor methodology.

In the preceding chapter, we looked at how the document planning component of an nlg system can produce a document plan that specifies both the content and overall structure of a document to be generated. The task of the microplanning component is to take such a document plan and refine it to produce a more detailed text specification that can be passed to the surface realisation component, which will produce a corresponding surface text.

The document plan leaves open a number of decisions about the eventual form of the text in the document to be generated. These finer-grained decisions are made by the microplanning component. In the nlg system architecture used in this book, the microplanner is concerned with

1. Lexicalisation. Choosing the particular words, syntactic constructs, and mark-up annotations used to communicate the information encoded in the document plan.

2. Aggregation. Deciding how much information should be communicated in each of the document's sentences.

3. Referring expression generation. Determining what phrases should be used to identify particular domain entities to the user.

The result is still not a text. The idiosyncrasies of syntax, morphology, and the target mark-up language must still be dealt with. However, once the text specification has been constructed, all substantive decisions have been made. It is the job of the surface realiser, discussed in Chapter 6, to then construct a document in accordance with these decisions.

We begin this chapter by presenting our view of the microplanning task in Section 5.1.

Before commencing our exploration of the technical content of work in natural language generation, in this chapter we take a step back and consider questions that arise when we look at putting nlg technology to work. We consider alternatives to nlg, and we discuss the circumstances under which it is appropriate to use nlg technology. We also look at requirements analysis and the evaluation and fielding of nlg systems. These topics are rarely discussed in the research literature but are of major significance when we are concerned with the construction of an operational natural language generation system.

Introduction

Research activity in natural language generation is sometimes carried out with relatively little attention being paid to how the fruits of the research might be transferred into a working environment. This approach may be entirely appropriate for work that focuses on research issues where even the questions to be asked are unclear; it can often be useful to abstract away from situations of use to clarify the underlying issues.

In this book, however, we also are concerned with how nlg technology can be used to build real working systems that are intended to be put into everyday use. This means that we have to think about issues such as the specification of system requirements and the problems that can arise in fielding systems.

In this chapter we look at the process of mapping an abstract text specification and its constituent phrase specifications into a surface text, made up of words, punctuation symbols, and mark-up annotations. This is known as surface realisation. Much of this discussion is centred around three software packages (kpml, surge, and realpro) which can often be used to carry out part of the mapping process.

Introduction

In the previous chapter, we saw how a text specification can be constructed from a document plan via the process of microplanning. The text specification provides a complete specification of the document to be generated but does not in itself constitute that document. It is a more abstract representation whose nature is suited to the kinds of manipulation required at earlier stages of the generation process. This representation needs to be mapped into a surface form, this being a sequence of words, punctuation symbols, and mark-up annotations which can be delivered to the document presentation system being used.

As described in Section 3.4, a text specification is a tree whose leaf nodes are phrase specifications and whose internal nodes show how phrases are grouped into document structures such as paragraphs and sections; internal nodes may also specify additional information about structures, such as section titles. Phrase specifications describe individual sentences; they may also describe sentence fragments in cases where these fragments are realised as orthographically separate elements (such as section titles or entries in an itemised list).

In this chapter and the two following, we turn to details of the components that make up the nlg system architecture we introduced in Chapter 3. Our concern in this chapter is with the component we have called the document planner.

The chapter is organised as follows. In Section 4.1, we give an overview of the task of document planning, including the inputs and outputs of this process; this is largely a recap of material introduced in Chapter 3. In Section 4.2, we look at domain modelling and the related task of message definition: Here we are concerned with the process of deciding how domain information should be represented for the purposes of natural language generation. In Sections 4.3 and 4.4, we turn to the two component tasks implicated in our view of document planning, these being content determination and document structuring. In each case we describe a number of different approaches that can be taken to the tasks. In Section 4.5, we look at how these tasks can be combined architecturally within a document planning module. The chapter ends with some pointers to further reading in Section 4.6.

Introduction

What Document Planning Is About

In the nlg system architecture we presented in Chapter 3, the document planner is responsible for deciding what information to communicate (this being the task of content determination) and determining how this information should be structured for presentation (this being the task of document structuring).

So far in this book we have considered the generation of natural language as if it were concerned with the production of text abstracted away from embodiment in any particular medium. This does not reflect reality, of course: When we are confronted with language, it is always embodied, whether in a speech stream, on a computer screen, on the page of a book, or on the back of a breakfast cereal packet. In this final chapter, we look beyond text generation and examine some of the issues that arise when we consider the generation of text contained within some medium.

Introduction

Linguistic content can be delivered to a reader or hearer in many ways. Consider, for example, a few of the ways in which a weather report might be presented to an audience:

• Email. When delivered as the body of a simple text-only email message, it might consist of nothing more than a sequence of words and punctuation symbols, with blank lines or indentations to indicate some of the structure in the text.

• Newspaper article. In this case it could include typographic elements, such as the use of bold and italic typefaces, and accompanying graphics, such as a weather map.

• Web page. As well as making use of typographic devices and graphical elements, in this case it could also include hypertext links to related information, such as the weather in neighbouring cities.

• Radio broadcast. Prosodic elements such as pauses and pitch changes might be used to communicate emphasis and structure.

• […]

The Purpose of This Book

This book describes natural language generation (nlg), which is a subfield of artificial intelligence and computational linguistics that is concerned with building computer software systems that can produce meaningful texts in English or other human languages from some underlying nonlinguistic representation of information. nlg systems use knowledge about language and the application domain to automatically produce documents, reports, help messages, and other kinds of texts.

As we enter the new millennium, work in natural language processing, and in particular natural language generation, is at an exciting stage in its development. The mid- to late 1990s have seen the emergence of the first fielded nlg applications, and the first software houses specialising in the development of nlg technology. At the time of writing, only a handful of systems are in everyday use, but many more are under development and should be fielded within the next few years. The growing interest in applications of the technology has also changed the nature of academic research in the field. More attention is now being paid to software engineering issues, and to using nlg within a wider document generation process that incorporates graphical elements and other realities of the Web-based information age such as hypertext links.

However, despite the growing interest in nlg in general and applied nlg in particular, it is often difficult for people who are not already knowledgeable in the field to obtain a comprehensive overview of what is involved in building a natural language generation system.

Natural language generation (nlg) is the subfield of artificial intelligence and computational linguistics that focuses on computer systems that can produce understandable texts in English or other human languages. Typically starting from some nonlinguistic representation of information as input, nlg systems use knowledge about language and the application domain to automatically produce documents, reports, explanations, help messages, and other kinds of texts.

nlg is both a fascinating area of research and an emerging technology with many real-world applications. As a research area, nlg brings a unique perspective on fundamental issues in artificial intelligence, cognitive science, and human–computer interaction. These include questions such as how linguistic and domain knowledge should be represented and reasoned with, what it means for a text to be well written, and how information is best communicated between machine and human. From a practical perspective, nlg technology is capable of partially automating routine document creation, removing much of the drudgery associated with such tasks. It is also being used in the research laboratory, and we expect soon in real applications, to present and explain complex information to people who do not have the background or time required to understand the raw data. In the longer term, nlg is also likely to play an important role in human–computer interfaces and will allow much richer interaction with machines than is possible today.

A piece of software as complex as a complete natural language generation system is unlikely to be constructed as a monolithic program. In this chapter, we introduce a particular architecture for nlg systems, by which we mean a specification of how the different types of processing are distributed across a number of component modules. As part of this architectural specification, we discuss how these modules interact with each other and we describe the data structures that are passed between the modules.

Introduction

Like other complex software systems, nlg systems are generally easier to build and debug if they are decomposed into distinct, well-defined, and easily-integrated modules. This is especially true if the software is being developed by a team rather than by one individual. Modularisation can also make it easier to reuse components amongst different applications and can make it easier to modify an application. Suppose, for example, we adopt a modularisation where one component is responsible for selecting the information content of a text and another is responsible for expressing this content in some natural language. Provided a well-defined interface between these components is specified, different teams or individuals can work on the two components independently. It may be possible to reuse the components (and in particular the second, less application-dependent component) independently of one another.

This chapter will not be strictly chronological and teleological. It is not an attempt to begin at a “beginning” and indicate a “development” to an ever-shifting present. It utilizes the concept of the “discursive formation,” where these formations were created within the political, social, and material conditions in South Africa. These broadly historical formations, identified by the distribution of “statements” within discourses, from approximately the sixteenth to the twentieth centuries, can be identified as follows: Europe meets Africa; the indigenization of language; colonization; literature as discourse; the phenomenon of a “minor literature”; modernism and postmodernism.

Europe meets Africa

The creation of a written literature in South Africa – literature as a nineteenth-century European construct – does not begin with the canonical “literary” text. It will take its representations from the navigation texts and travel journals of those who documented the first meetings and confrontations in the contact zone between the indigene, the Portuguese and Dutch seamen and explorers. The travel discourses of the first Portuguese and Dutch navigators who sailed around the southernmost point of Africa, Bartolomeu Dias (1487–88), Vasco da Gama (1497–99), and Jan Huygen van Linschoten (1579–92) (see Axelson 1988: 1–8; Itinerario Voyage 1934), which record what seems to be the first significant impressions of the people and landscape of southern Africa, have to be read not only for their content as texts describing what they saw and experienced. They are also representations in language, limited as instruments of representation; but also powerful as textual creations constructing images of the other people as wild, barbaric, dirty, stupid, and untrustworthy.