Metaphor in end-of-life care (MELC)

Detailed Background to MELC


Metaphor is recognised in the human, social and cognitive sciences as a ubiquitous cognitive phenomenon and powerful communicative tool, which can reflect conventional and implicit ways of thinking, and which can help to overcome communicative challenges, as well as contribute to communicative problems. Our research provides a detailed account of the ways in which members of different stakeholder groups employ metaphor to frame their experiences of, and approach to, end-of-life care, and particularly to express their attitudes, emotions and evaluations, perceptions of themselves, others and their mutual relationships, and views on personal and professional issues. Our research also sheds light on the ways in which the use of metaphor can reflect or contribute to anxiety and misunderstanding, especially when topics are personal, sensitive or perceived as taboo, and when asymmetries of power and expertise exist in the relevant domain of experience.

Over the last three decades, scholars working within 'Cognitive Metaphor Theory' (hereafter CMT) have drawn attention to the presence of pervasive conventional metaphorical patterns in language, involving linguistic expressions such as 'She died after a long battle with cancer' or 'She passed away in her sleep'. These metaphorical patterns have been seen as evidence for the existence of 'conceptual metaphors', namely, conventional patterns of thought that involve systematic sets of correspondences across domains in the conceptual structure of speakers of a language. Thus, the metaphorical expressions mentioned above have been explained as linguistic realizations of, respectively, the conventional conceptual metaphors A DIFFICULT ENTERPRISE IS A WAR and DEATH IS DEPARTURE/GOING AWAY. Within these conceptual metaphors, WAR and DEPARTURE are 'source' domains and DIFFICULT ENTERPRISES and DEATH are 'target' domains (Lakoff and Johnson 1980, 1999). The same conceptual metaphors may be realized by 'novel' as opposed to 'conventional' metaphorical expressions, as is the case with the underlined expressions in the following extract in which a cancer sufferer describes his treatment in terms of war: 'I have surrounded myself by barbed wire, land mines, and several squads of infantry, and we are ready to take on all comers' (quoted in Reisfield and Wilson 2004: 4025)

Target domains typically correspond to areas of experience that are relatively abstract, complex, unfamiliar, subjective or poorly delineated, such as emotions, life, death, and time. In contrast, source domains typically correspond to concrete, simple, familiar, physical and well delineated experiences, such as motion and concrete objects. The choice of metaphor affects the way in which the target domain can be most easily thought and talked about, including by means of narratives that exploit aspects of the source domain (see Musolff 2006). For example, when applied to illness, the metaphor A DIFFICULT ENTERPRISE IS A WAR highlights the way in which being ill may involve strength, perseverance, endurance and heroism, and facilitates an understanding of recovering health as a victory and of not recovering or dying as a defeat. In contrast, this metaphor backgrounds other ways of making sense of health problems, such as accepting and living with illness. However, the shortcomings of a particular metaphor may be overcome by replacing it or using it alongside an alternative metaphor. For example, the metaphorical construction of being ill as a journey can highlight both progress and lack of progress, as well as routine, companionship, and so on. Indeed, recent policy documents on cancer in particular suggest a deliberate shift away from military metaphors towards JOURNEY metaphors, such as the 'cancer journey' and clinical 'pathways' mentioned above.

Metaphor in health communication

Illness is an individual, personal state of being, which is associated with physical discomfort or pain, and feelings of anxiety, fear, isolation, and, potentially, shame. As such, it belongs to the kind of complex, subjective, and poorly delineated experiences that tend to be conventionally verbalised and conceptualised through metaphor. Crucially, however, sufferers, carers and medical professionals need to interact and cooperate in the process of care and treatment, and metaphor may be used more or less successfully, sensitively, and harmoniously within these interactions.

Susan Sontag famously exposed the potentially damaging role of metaphor in discourses surrounding illness, particularly in relation to cancer (Sontag 1979) and AIDS (Sontag 1988). Sontag claimed that WAR metaphors contribute to the demonization of illnesses such as cancer, and to an unnecessary increase in patients' feelings of fear, helplessness and isolation. She therefore argued for the elimination of metaphor from the discourses surrounding illness, and looked forward to a time when medical advances will make metaphor unnecessary. While Sontag's contribution remains highly influential, scholars in a range of areas recognize that metaphor enables both professionals and patients to construct accounts of illness that provide meaning, for example by means of the notions of a heroic battle with cancer or of a journey of (self-)discovery. A number of studies by medical professionals have described metaphor both as a useful resource and as a potential danger in communication about illness. Metaphor is seen as a resource insofar as it enables sufferers to express and share their experiences, and professionals to clarify various aspects of diseases and treatments. It is a danger when, as Sontag showed, it contributes to representations of illness that are demoralising and confusing for sufferers and those close to them. What tends to be advocated, therefore, is not the elimination of metaphor, but a more conscious, sensitive and effective use of metaphor, especially on the part of professionals interacting with patients (e.g. Canter 1988, Czechmeister 1994, Reisfield and Wilson 2004).

The primary aim of the MELC project is to build on this earlier work, and on relevant work on metaphors for emotions, life, death and time (e.g. Lakoff and Johnson 1999, Kövecses 2000), in order to provide an account of the use of metaphor on the part of members of different stakeholder groups in end-of-life care, and of its implications for a better understanding of the experiences, needs and challenges faced by these different groups.

Corpus-based research on metaphor

In recent years, some researchers have started to use electronic corpora and corpus analysis software for the investigation of metaphorical patterns (e.g. Charteris-Black 2004, Koller 2004, Deignan 2005, Semino 2005, Stefanowitsch and Gries 2006). These scholars employ corpus-based methods primarily in order to (a) test the claims of CMT by conducting systematic investigations of metaphorical expressions in general-purpose corpora (e.g. Deignan 2005), or (b) conduct specific investigations of patterns of metaphorical expressions in particular genres and discourses, and their implications for, for example, dominant ideologies (e.g. Koller 2004). These studies have shown how the systematic analysis of metaphorical patterns in naturally-occurring data raises further questions and leads to further insights into metaphor as a linguistic and cognitive phenomenon.

The exploitation of electronic corpora for metaphor research is constrained, however, by the fact that the identification of metaphorical expressions in texts has not yet been successfully automated, despite some promising attempts (e.g. Mason 2004, Berber Sardinha 2010). Hence, existing studies have relied on a combination of traditional 'manual' analysis with the concordancing of selected metaphorical expressions in electronic data (i.e. the use of software that provides all instances of individual search strings in the corpus under investigation) (e.g. Skorczynska and Deignan 2006). However, this methodology only allows researchers to find further instances of previously identified expressions.

In order to overcome these limitations, we employ an adapted version of the UCREL Semantic Analysis System (USAS) tagger (Rayson et al. 2004) embedded within the Wmatrix tool (, which was developed by one of the CIs (Rayson). This tool can be employed to search for expressions belonging to semantic fields that are likely to correspond to metaphorical source domains, and hence to identify potential metaphorical expressions in large data sets without being restricted to pre-established lists of lexical items.

Research Questions


  • A. How do members of different stakeholders groups (health professionals, patients and unpaid family carers) use metaphor to talk about their experiences, attitudes and expectations of end-of-life care (e.g. palliative treatment, preparations for dying, etc)?
  • B. What does the use of metaphor by these stakeholder groups suggest about (a) the experiences and needs of the members of these groups and their mutual relationships, and (b) the nature of metaphor as a linguistic and cognitive phenomenon?


  • 1. How do health professionals use metaphor to talk about end-of-life care?
  • 1.1. What metaphorical expressions and patterns can be identified in: (a) interviews with health professionals, and (b) contributions to health professionals' online fora on the topic of end-of-life care?
  • 2. How do patients approaching end of life use metaphor to talk about end-of-life care?
  • 2.1. What metaphorical expressions and patterns can be identified in: (a) interviews with patients approaching the end of life and (b) contributions to online fora by patients approaching the end of life?
  • 3. How do unpaid family carers use metaphor to talk about end-of-life care?
  • 3.1. What metaphorical expressions and patterns can be identified in: (a) interviews with unpaid family carers on the topic of end-of-life care and (b) contributions to unpaid family carers' online fora on the topic of end-of-life care?
  • 4. What does the use of metaphor suggest concerning:
  • 4.1. the experiences, attitudes and expectations of the members of each different group?
  • 4.2. differences and similarities in experiences, views, and attitudes within and across the different groups?
  • 4.3. the influence of recent policy changes on the language used by different stakeholder groups?
  • 4.4. the consultancy for, and training of, health professionals?
  • 5. What are the implications of our findings for metaphor theory?
  • 5.1. How do the metaphorical expressions realising different source domains vary in terms of their linguistic characteristics (type-token ratio, part of speech distribution, conventionality vs. novelty) and their frequencies and functions, both within and across each section of the corpus?
  • 5.2. How do the patterns we have identified relate to the findings of earlier studies of the linguistic manifestations of metaphor in different genres and the discourses with which they are associated?

Research Methods

The most innovative aspect of our methodology is the exploitation of a technique developed and tested in the earlier stages of our research, and instantiated as a tool within the Wmatrix environment. In most earlier approaches to metaphor in large corpora, corpus searches have been limited to running concordances for selected sets of word forms. In our approach, the linguistic knowledge in the semantic tagging system (namely, its lexicon) is used to identify a much more comprehensive set of potentially metaphorical expressions within a given source domain.

This technique, which we refer to as broad-sweep searching, is a means by which computational analysis can be used to assist, channel and direct manual analysis of metaphorical patterns. This makes it possible for much larger quantities of data to be handled. However, 'manual' analysis is needed both before and after the computational analysis, in order to identify in a sample from the corpus a preliminary set of candidate semantic domains to be further investigated via the semantic annotation tool, and to identify and tag the genuinely metaphorical instances in the output of the tool.

Our method for distinguishing between metaphorical and non-metaphorical expressions is based on the procedure described in Pragglejaz Group (2007). We operationalise the distinction between conventional and novel metaphorical expressions by consulting corpus-based dictionaries of English (e.g. the Macmillan English Dictionary): we regard as conventional those metaphorical expressions whose metaphorical senses are included in contemporary corpus-based dictionaries; other metaphorical expressions are regarded as novel.

The corpus

Our corpus contains approximately 1.5 million words, broken down into three sections of approximately 500,000 words each, corresponding to three main stakeholder groups in end-of-life care: health professionals, patients, and unpaid family carers. Each section of the corpus includes transcriptions of ca. 15 semi-structured interviews of ca. 45 minutes with members of each group (approximately 100,000 words for each group), and data harvested from online fora dedicated to each group (approximately 400,000 words for each group) (see Seale et al. 2010's discussion of interviews and online fora as data in qualitative health research).

For a list of references see here

ESRC logo Lancaster University logo International Observatory on End of Life Care UCREL research centre

This page last modified on Friday 16 November 2012 at 5:01 pm .