Towards a Dynamic Approach to Analysis

Lakoff and Johnson do not present a method for analyzing metaphors as such. Others in turn have done this. Below, three different analysis approaches will be presented and put into a framework of top-down and bottom-up analyses provided by Krenmayr (2013). First, I present the Metaphor Identification Process (MIP) and then the Discourse Dynamics Approach (DDA). Last, I present a merge of the two in Jensen and Cuffari’s combination of metaphor analysis and distributed cognition.

The approach taken by Steen (2008, 2010) as well as the Pragglejaz Group (2007) and further developed in the MIP VU method (Steen, 2010) holds two huge differences from the approaches seen above. The purpose is to investigate metaphors from an inductive rather than deductive angle, and there is a differentiation between conceptual and linguistic metaphors.

Lakoff and Johnson as well as the other developers of metaphor theory described above test hypotheses and thus place the expectations connected to a certain metaphor into the analysis that verify or falsify the hypothesis. The aim of Steen and others is to go beyond that and develop a method for a much more systematic and inductive way of investigation.

Steen relates to the problem of comparison or categorization by making a distinction be-tween deliberate and non-deliberate metaphors. He claims that deliberate metaphors are

processed as comparison and thus in a linguistic understanding are processed metaphorically whereas the non-deliberate metaphors are processed as categorization and not metaphor-ically in linguistic terms. Nevertheless, Steen still counts non-deliberate metaphors as metaphors.

A deliberate or communicative metaphor is defined as a metaphor used to change the lis-tener’s perception or view of something. The non-deliberate metaphors are either linguistic or conceptual, leading to three types of metaphors:

1. Language metaphors (linguistic form, poles being a metaphor and a non-metaphor) 2. Thought metaphors (conceptual structure, poles being a conventional and a novel

metaphor)

3. Communicative metaphors (Communicative function, poles being deliberate and non deliberate).

The three dimensions are not interdependent as in e.g. the Career of metaphor theory where metaphors are either novel, conventional or dead. They are different dimensions of what a metaphor is and can do.

By using the three dimensional model, the different kinds of metaphors defined and discussed above become dimensions on the same fundamental understanding of a metaphor as a cross domain mapping between an abstract target domain and a well-known source domain. It buries the Career of metaphor theory and defines the discussion of comparison versus categorization as a question of dimensions rather than of completely different views of metaphors.

Instead of doing the direct cross domain mapping, the method here is much more structured and schematic when analyzing a text corpus. The Pragglejaz group (2007) offers a four-step model to inductively investigate simile or comparison between target and source domain:

1. Read the text-discourse to establish a general understanding 2. Determine the lexical units in the test-discourse

a) Establish contextual meaning for each lexical unit

b) Establish if there is a more basic, contemporary and bodily meaning c) Decide if the contextual meaning differs from the basic meaning 3. If yes, mark unit as metaphorical.

The MIP-VU procedure (Steen 2010) is even more rigid and thorough in establishing verifiable and reproducible results in metaphor analysis. Going through these four-five steps is a movement from what is potentially metaphorical to establishing the metaphor in context. By use of a dictionary, basic and contextual meaning is established, and if they differ, the use is metaphorical. This approach in a modified version has been used in the analysis presented in chapter 7 as well as in article 1.

A different analysis approach to metaphors is found in Cameron et al (Cameron & Deignan, 2006; Cameron et al., 2009). This approach is minded on conversational data and is thus immediately suitable for the purposes of my study.

30

The purpose of the DDA approach is:

... to investigate the intricate web of sense-making and sense-creating that happens when people engage in spontaneous discourse” (Cameron et al., 2009, p. 83).

What Cameron and colleagues propose is an analysis focusing more on the behavior of metaphor in conversations than on the specific words and their basic and contextual meaning as is the case in the MIP and MIP-VU. Analyzing conversation by coding conversations and by use of the metaphors extracting how metaphors “uncover people’s ideas, attitudes and values” (ibid, p. 64) is less focused on the linguistic difference between basic and contextual meaning and more focused on how metaphors emerge and travel through conversations.

As my data consist of conversations, a combination of the two approaches was used. All conversations were coded in the tradition of Cameron and colleagues, but when establishing whether a certain word could be considered a metaphor, the MIP-procedure was applied.

The two approaches might initially seem very different — one is very methodical and word-focused, and the other is more concerned with sense-making, but combined they form an approach to understanding metaphors in discourse which is more nuanced in the case of conversational data than either can provide alone. However, in coding the conversations, it became more and more evident that more modes contributed to the emergence and understanding of metaphors for knowledge, and more layers had to be added.

Jensen and Cuffari present what they call an “experience-oriented methodology” (Jensen

& Cuffari, 2014, p. 280). This is very much inspired by DDA but removes the focus from metaphors in words only. They regard conversation as a coupled action where participants co-create metaphors as distributed cognitive systems like it has been suggested by Hutchins (1995) as mentioned above and by Gibbs and Santa Cruz (2012). This approach does not focus exclusively on the words in conversations but on the whole event of the conversation.

They look for behavioral patterns of metaphoricity or a doubleness in experience. This is done by coding words with images. The result of this approach is much more interactional and does not so much relate to the specific metaphors in words. However, in understanding metaphors as emergent in conversations, this multimodal approach is very appealing and will be touched upon again in chapter 3.4 regarding multimodality.

To sum up and put in perspective this often applied approach to CMT, Krenmayr (2013) presents two approaches to metaphor analysis in natural conversation. She names the two the top-down and the bottom-up approach. The MIP procedure as well as the DDA approach and the multimodal approach would fall in the bottom-up category whereas the deductive approach would fit the top-down category. Both MIP and DDA are inductive approaches, and as described above, they are not in line with the deductive approach of e.g.

Lakoff and Johnson. In some senses, this approach to metaphor analysis is therefore an expansion of the understanding of CMT but nevertheless in close relation to the theory.

The real difference between the two approaches, though, is that one is focusing on linguistic and the other on the conceptual level. According to Krenmayr, these two are easily confused, and it is important for the researcher to be clear on when linguistic metaphors are analyzed and when conceptual metaphors are the object. This is in line with the three-dimensional model but less relevant if what is studied is the degree of metaphoricity in multiple modes.

I will return to this in chapter 3.4 below.