Natural Language Processing

Some Further Information Concerning the Research Fellow Position
in the project entitled

Metaphor and Metonymy: Addressing a Debate and a Neglected Problem

funded by the Leverhulme Trust



The University of Birmingham

The University of Birmingham is a major international centre of academic excellence. It is a member of the Russell Group (of top research universities in the UK) and belongs to the international network Universitas 21. It is tenth and eleventh among all UK universities according to the Times Higher Educational World University Rankings and the Jiao Tong ranking respectively.

The University was founded in 1900 and is now one of the largest in the UK, offering degrees and undertaking research across a wide range of disciplines from Education to Medicine and from Engineering to Law. It is currently proceeding with an extensive capital development programme. This positive financial position is almost unique in the UK Higher Education sector.

The University provides a beneficial academic setting for the Leverhulme Trust project in figurative language. Many researchers across the university (e.g., in English, Education, Psychology) are interested in aspects of the field of metaphor, and the Philosophy department contains researchers in the philosophy of language and mind. In addition the University has played an important role in the development of corpus linguistics.

For further information about the University see its website.



The Location

One of the University's greatest general assets is its main campus in Edgbaston. It offers its community of over 25,000 students and 6,000 staff an attractive environment in which to study and work. It is only three miles from the centre of a major European city and yet is set amongst green and leafy parkland which is largely pedestrianised, and provides a beautiful and pleasant backdrop for imposing Victorian redbrick buildings as well as some striking modern architecture.

The city centre has in recent years developed some of the most attractive city centre facilities in the country, notably the the new Bullring shopping complex and the canal-based Brindley Place area.

Birmingham has a vibrant popular culture (clubs, concerts, regular festivals, etc., and facilities such as the National Indoor Arena), excellent restaurants, and a strong international profile in the fine arts (classical music, theatre, ballet, visual art). In particular the City of Birmingham Symphony Orchestra (CBSO) and the Birmingham Royal Ballet are internationally renowned. The CBSO is centred at the Symphony Hall within the spectacular International Convention Centre. The University itself has, in its Barber Institute, and important centre for visual art and muscial performance.

Birmingham has an international airport linking it direct to many important cities and hubs across the world. It is now less than one hour and a half hours by train from London, and has good rail, bus and road connections also to all corners of the UK, making it often the meeting place of choice for academic and other bodies. It is close to beautiful countryside and to other major cultural centres such as Oxford and Stratford upon Avon (the latter providing fine contemporary theatre as well as being the main hub worldwide for Shakespearean theatre).



The School of Computer Science

The School is a minute's walk from the University railway station on the Edgbaston campus, connecting quickly and frequently to the city centre (in about five minutes, with a train every ten minutes), and to surrounding towns.

The School is in a modern and airy building, built in 2000, providing one of the most attractive and companionable working environments on campus. There are offices holding between two to four research fellows, and these offices are adjacent to the offices of lecturers and professors. There are meeting rooms providing attractive space and easy opportunities for spontaneous interaction as well as planned discussions. The computer network within the School is run by local computer officers who cater for the specialized needs of computer scientists.

The School has an intense research-seminar culture. As well as a broad-coverage weekly seminar series, there are weekly series in CS Theory and in Artificial Intelligence and Natural Computation.

With major new investment recently awarded by the University, the School is setting up two new major cross-disciplinary research centres: the Human-Computer Interaction centre (joint with the department of Electrical, Electronic and Computer Engineering) and the centre in Computational Neuroscience and Cognitive Robotics (joint with the School of Psychology).

For further information about the School, including its research, see its website.



The Leverhulme Trust Project in Figurative Language

The Leverhulme Trust funding is prestigious, and the Trust concentrates its funding on particularly adventurous and boundary-crossing research.

The following notes add to the information given in the advertisement of the Research Fellowship on the project, and is largely taken from the grant proposal. A more detailed description of the envisaged work is available on request (to John Barnden at jab@cs.bham.ac.uk). Also, "Avenue 1" of the project arises directly from the following paper:


The Context for the Research

Metaphor and other "tropes" (types of figurative language) are an important research focus in many fields: Linguistics, Literature, Philosophy, Psychology, Psychiatry, Education, Business, Politics, etc. Many investigators have recognized metaphor to be pivotal, not just important, in the study of meaning and thought. The tropes are important within Artificial Intelligence because AI systems need ultimately to communicate naturally with people. Tropes are important not just in poetry, but also in mundane discourse: ordinary conversation, news articles, popular novels, advertisements, etc.

The research will be along two Avenues:-

Research Avenue 1

Researchers have intensely debated how metaphor differs from another pervasive trope, metonymy. Metonymy consists of talking about something through talking about some connected matter without making the link explicit. Consider "John listened to Bach." If this is about John listening to some Bach music, a link from Bach to his music is implicitly operative. Metonymy may initially look distinct from metaphor but it has proven difficult to distinguish them rigorously. The current proposal takes a radical, provocative stance on this issue. Researchers have appealed to various alleged distinctions, such as:

  • metaphor involves similarity whereas metonymy involves perceived "contiguity" (as between Bach and his music)

  • concepts fall into different compartments, and metaphor crosses between compartments (e.g., in viewing anger as heat, we cross between a concept in a physical-world compartment and a concept in an emotions compartment) whereas metonymy stays within a compartment (e.g., within the music world)

Adding to other researchers' criticisms of various such grounds, I have shown that the metaphor/metonymy differentiation problem is deeper and more extensive than previously thought. For example, some important types of contiguity themselves involve similarity. Consequently, I propose that metaphor and metonymy are just loose, intuitive notions that should not be the main focus of technical discussion. We should analyze linguistic expressions more fundamentally, as involving varying degrees and types of similarity, contiguity, compartment-crossing, etc., taking these just as dimensions of variation, not as differentiators. Metaphor and metonymy can both involve some similarity, contiguity, etc. They may tend to interact somewhat differently with the dimensions, but there are no firm, definable boundaries, and technical reality is better described at the level of the dimensions.

A (Largely) Neglected Problem: Research Avenue 2

A few philosophers have studied the embedding of metaphors (etc.) in mental-state contexts, as in "Romeo thinks that Juliet is the sun", whereas most researchers are content with studying unembedded metaphors -- say, "Juliet is the sun" by itself. Such embedding is technically challenging and disruptive for both figurative-language theory and general theory of language meaning. But the problem has largely been neglected, partly because of traditional divisions within disciplines. And mental-state contexts are by themselves a major source of technical anxiety across disciplines, just as metaphor by itself is. However, it is imperative to break down the artificial division.


Specific Objectives of the Project

AVENUE 1:

  • study metaphor and metonymy further to challenge and enrich the dimensional deconstruction

  • expand it to address figurative simile, hyperbole and understatement

  • rework Philosophical claims about metaphor as necessary to accommodate the deconstruction

  • revise an existing AI theory (called ATT-Meta) of the process of metaphor understanding (project description and papers).

AVENUE 2:

  • challenge or suitably refine some recent claims in Philosophy about mentally-embedded metaphor and the ramifications concerning language meaning in general

  • explore the consequences, for theories of meaning in Philosophy, Linguistics and AI, of the following novel "metaphorization" hypothesis: in interpreting "Mike thinks that Y" where Y is NOT metaphorically expressed, it may nevertheless be appropriate to impute to Mike a metaphorical view of what Y expresses

  • develop the preliminary treatment of mentally-embedded metaphor in the ATT-Meta theory

THE AVENUES' INTERSECTION:

  • generalize and refine Avenue 2 to account for the lack of a distinct boundary around metaphor and to couch issues explicitly in terms of the underlying dimensions themselves

  • refine the Avenue 1 analysis because, e.g., differing beliefs about the world need to be accounted for in defining what, say, contiguity is.


Significance and Originality

The project will illuminate one of the most puzzling and profound features of language, namely figurative expression, and will fundamentally affect the analysis of language about mental states, another long-studied, pivotal topic in Philosophy, Linguistics, AI and elsewhere. The study will also affect the broad question of what language meaning is; in particular, the study tends towards a highly liberal view, flouting certain restrictions that others have imposed.

Avenue 1's effect on figurative language studies is potentially great. Some researchers have discussed fuzziness of tropes but none have previously proposed a multidimensional deconstruction. The account has highly original specific aspects such as analysing contiguity as sometimes involving similarity.

Avenue 2 marries two areas of study (metaphor; mental-state reports) that have largely been artificially separated. As one benefit, the semantic "metaphorization" of non-metaphorical expressions (see second Avenue 2 Objective above) constitutes a radical new view about the involvement of metaphor in meaning, and suggests profound advances as regards "nested" mental states ("Mike believes that Sally hopes that ... ") because inner mental states (e.g. Sally's hope) may be metaphorized.

Elucidating the interactions of Avenues 1 and 2 will itself be a significant contribution.

Advances made will be important for theoretical, and in some cases practical, development in many academic fields and application areas concerned with figurative language. Avenue 1 has practical importance for management and empirical study of figurative language usage in areas such as Psychiatry, Education, Applied Linguistics, Business Studies, and Politics. For example, Applied Linguists are increasingly interested in how to address figurative language in language teaching. The project should also inspire revised forms of Psychological experiment concerning people's understanding of tropes, for example to tease apart the individual dimensions and to clarify the effects of embedding (but the project will not itself conduct experiments).