1 In this article, I explore Amos’s self-construction (see Appendix) as it is identified, described, and interpreted via a metaphor-oriented positioning analysis2 (MPA) presented and evaluated.

2 The metaphor analysis presented in this article is inspired by a functionalist approach 3 to discourse (Schiffrin, 1994) that defines discourse as language-in-action, and emphasizes that discourse analysis “requires attention to both language and action” (Blommaert, 2005, p. 2); (e.g., the narrator produces a metaphorical cluster to position himself as a valuable kibbutz member).

3 The functionalist approach (Schiffrin, 1994) espoused in this article also foregrounds the centrality of the context in which language is produced (Linell, 1998). For example, in order to analyze Amos's life story, I considered the immediate interview setting as well as other information regarding his health that was supplied by the interviewer. In addition, the approach emphasizes the importance of self-building language resources including metaphors (Cameron, 2009; Semino, 2008) in the study of narrative discourse. These are discourse guides indicating who interlocutors are speaking as (Malone 1997); (e.g., as an individual “I,” as a collective “we,” as an opposing collective “they,” or as a distanced self “you”).

4 Following Georgakopoulou (1997), language resources are not defined as preconceived lists of linguistic devices, but rather attention is paid to their specific functions in the context in which they are produced.

5 The analysis also aligns itself with discursive psychology (Hepburn & Wiggins, 2007)—an approach that often employs conversation analytic tools (Heritage & Clayman, 2010) in order to explore how psychological phenomena such as self-construction are interactively accomplished rather than being regarded as a priori properties of the individual (Widdicombe, 1998).

6 The focus on sequences of speech turns is considered a central tool employed by conversation analysts (Schegloff, 2007) when they study human interaction. Accordingly, I view Amos’s life story as a turn in the interaction between him and the interviewer. This turn is oriented to the interviewer’s request to unfold a life story. In the ensuing analysis, I will show that although the interviewer's turns do not appear in the text, they constitute the context that, by and large, influences the content and form of Amos' life story.

7 Discursive psychologists associate self-construction with positioning (Hepburn & Wiggins, 2007)—a central social action accomplished in discourse “whereby selves are located in conversations as observably and subjectively coherent [emphasis added] participants in jointly produced story lines” (Davies & Harré, 1990, p. 40).

8 This definition foregrounds two characteristics of positioning. First, it stresses that positioning involves narrators' self-construction in interaction in relation to others rather than being regarded as a priori properties of the individual (Widdicombe, 1998). The second characteristic associates positioning with coherent narration. Life stories, the narrative genre on which this special issue focuses, are regarded as coherent when narrators organize them temporally, causally, thematically, and morally (McAdams, 2006).

9 Current publications focusing on the experience and self of the troubled and the ill (Freeman, 2010; Hydén & Brockmeier, 2008; Hyvärinen, Hydén, Saarenheimo, & Tamboukou, 2010; Kupferberg, Gilat, Dahan, & Doron, 2012, 2013) challenge and problematize the emphasis on coherent narration in contexts where the expression of self is undermined by illness or trouble.

10 This is the context in which Amos's life story is told. Therefore, it is intriguing to explore how the narrator, an 85-year-old man who suffered a stroke 15 years prior to the interview, positions himself vis-à-vis the interviewer as well as significant others in the narrated past and possible future landscapes (Ochs, 1994).

11 In the first two sections of the article, the theoretical framework underlying MPA is presented. In the third section, the methodological and methodical frameworks are reviewed. In the fourth section, the analysis is presented, and in the discussion, the insights gleaned from the analysis and the contribution of the positioning analysis are described and evaluated.

12 The metaphor-oriented positioning analysis (MPA) presented in this article was initially developed for the study of naturally-occurring (Speer, 2007) face-to-face, telephone and computer-mediated troubled communication (Kupferberg, 2008, 2010a, 2010b; Kupferberg & Green, 1998, 2005), and later adapted to narrative interviewing, large corpora of stories, and non-narrative discourse (see overview in Kupferberg, 2010b).

13 MPA foregrounds the centrality of narrative time (see overview in Freeman, 1998), a theoretical construct defined as “a back-and-forth movement between the past and the present that furthermore relates to the future, even if it might not always be present” (Brockmeier, 2000, p. 54). Accordingly, narrative time enables humans to overcome the everlasting linearity of chronological time by focusing on the complexities of the past and the possibilities offered by the future while they are engaged in the present ongoing conversation (Kupferberg & Green, 2005). In other words, although time passes, humans can go back to the past and the future in their thinking.The present moment constitutes a “workshop” in which humans interactionally attempt to reach coherence or agreement on the meaning of their past and future.

14 Following Bublitz (1999), MPA defines coherence as an action that is co-constructed with other interlocutors rather than regarding it as an accomplishment of the narrator, as McAdams's (2006) definition that was presented earlier implies (for detailed description of coherence see overview in Kupferberg & Green, 2005; Kupferberg et al., 2012).

15 MPA is also inspired by Bamberg’s tenet (1997a, 2004, 2006) emphasizing that, when humans tell personal stories, they often position or locate themselves in certain ways not only in relation to others who are present in the interaction, but also in relation to significant non-present others in the narrated past. Alternatively, interlocutors are sometimes positioned by others in the ongoing interaction, or by the context in which they live (Bamberg, 2006; van Langenhove & Harré, 1999).

16 Assuming that human mental life actually moves from the present to the past and the future, then it is interesting to examine closely how narrators position themselves in relation to others at each of these levels in conversation. For example, Amos positions himself at the end of the interview in relation to the interviewer at the present level (Level 1) “And that’s that. About myself. What else do you want to hear? Interesting?” (59-60). He also positions himself vis-à-vis other kibbutz members at the past level (Level 2) when he talks about the plant that he established: “And in the beginning it limped along a bit. And then (they) actually began to run after me. Why did you create this white elephant and why that (31-33).” As the level analysis will show, Amos does not position himself at the future level (Level 3).

17 The author's current studies (Kupferberg & Gilat, 2012; Kupferberg et al., 2012; Kupferberg & Hess, 2013) further emphasize that narrators’ advertent or inadvertent positioning at one level or another is self-revealing. The first study (Kupferberg & Gilat, 2012) combines qualitative and quantitative methods and shows that troubled narrators of varying ages frequently narrate their problems (Level 2) and develop interpersonal relations (Level 1) when they seek help in a computer-mediated forum for mental help. However, suicidal narrators who participated in the same forum avoid narration (and consequently self-construction) as well as interpersonal communication and merely “invest” in the construction of a short-term future (Level 3) when they intend to escape from their problems by committing suicide.

18 In another study that focuses on the life story of a schizophrenic inpatient (Kupferberg et al., 2012), the level analysis shows when and how the narrator loses and regains her ability to position herself coherently in discourse. At the level of interpersonal communication (Level 1), the narrator accomplishes the social discursive tasks of constructing different but coherent dimensions of herself in relation to the interviewer. However, when she positions herself in relation to significant others in her tormented past world (Level 2), or attempts to plan future actions (Level 3), she often fails to present coherent dimensions of herself and at times explicitly wonders who she is.

19 In the third study (Kupferberg & Hess, 2013), qualitative and quantitative level analyses show that adults with visual impairments and blindness (VIB) using a computer-mediated forum for people with VIB frequently used Level 1 to co-construct inter-personal communication with other forum members. At this level, they co-constructed their thoughts and emotions with other forum members. They also resorted to Level 2, which enabled the participants to construct the meaning of relevant past experience by means of varying specific and generic stories that they presented in the forum. Finally, the use of Level 3 showed that the participants occasionally co- constructed a future that included proposals and plans that could improve the conditions of their lives by changing norms and regulations. The three studies show that narrators’ advertent or inadvertent positioning at one level or another is self-revealing and may shed light on the communicative behavior of individuals or groups of people.

20 MPA's reliance on metaphors and other language resources constitutes its unique feature. We define metaphor as “the phenomenon whereby we talk and, potentially, think about something in terms of something else" (Semino, 2008, p. 1). Following MPA, one-word and multi-word metaphors are defined as central self-displaying positioning devices (Kupferberg & Vardi-Rath, 2012; Kupferberg & Green, 1998, 2005, 2008a, 2008b) that often summarize the gist of the story. In this way, narrators sometimes produce detailed (i.e., narrative) and succinct (i.e., metaphorical) versions of themselves (Kupferberg & Green, 1998, 2005).

21 In addition, metaphorical clusters “tend to cluster at certain points in the talk” (Cameron, 2009, p. 200) which constitute “critical moments in the discourse” (Cameron & Stelma, 2004, p. 33) when narrators perform meaningful intensive work (Kupferberg & Gilat, 2012; Kupferberg & Green, 2008a, 2008b; Kupferberg & Tabak, in press) as they interact with other interlocutors. To show how clusters function, one has to examine them attentively (Cameron & Stelma, 2004).

22 Metaphors do not work in isolation. Interlocutors often integrate them with other language resources such as syntactic structures, lexical items, constructed dialogue, rhetorical questions, and pronouns when they attempt to make meaning in the ongoing conversation (Kupferberg & Green, 2005).

23 MPA espouses a discourse-oriented approach 4 to metaphor that underscores the centrality of metaphors in the study of narrative and non- narrative discourse (Cameron, 2007, 2009; Kupferberg & Green, 2005; Semino, 2008). This approach differs from Conceptual Metaphor Theory (CMT), which defines metaphors as culturally-shaped (Kövecses, 2005) mental representations that determine the use of linguistic metaphors produced in human communication (Gibbs, 1994; Lakoff, 1993; Lakoff & Johnson, 1980, 1999).

24 CMT proponents emphasize that metaphors enable humans to conceptualize abstract and inexpressible target topics in terms of more familiar source domains (Lakoff & Johnson, 1980) (e.g., LIFE IS A ROAD is an example of a conceptual metaphor. The word “road” is the source domain used to conceptualize the abstract target domain “life”). Previous studies (Cameron & Low, 2011; Cameron & Stelma, 2004; Kupferberg & Green, 2008a; Semino, 2008) show that CMT-based analyses sometimes fail to account for the complexity that is characteristic of naturally-occurring discourse.

25 Proponents of the discourse-oriented approach to metaphor explore linguistic metaphors in interactional discourse in situ and in vivo rather than relying on predetermined conceptual metaphors. Linguistic metaphors are defined as a one-word or multi-word unit in which the contextual meaning of a word in the text is different from the basic dictionary meaning of the same word or unit and should be understood in comparison to it (Pragglejaz Group, 2007). For example, Amos conceptualizes the meaning of his army service in terms of “drifting” (15). The basic dictionary meaning of “drifting” (i.e., “to become driven or carried along as by a current of water, wind, or air,” Merriam Webster Dictionary) is used by Amos in order to conceptualize the experience of serving in the army. Therefore, we can conclude that “drifting” in the context of Amos's story is metaphorical.

26 Metaphor researchers also distinguish between deliberate and conventional metaphors. "Deliberate metaphors are different from conventional ones, which are typically produced automatically and thoughtlessly, something that speakers and listeners, authors and readers , tacitly recognize when they engage in metaphoric discourse" (Gibbs, 2011, p.26).

27 Metaphors become conventional “when the relevant metaphorical meaning has become lexicalized, so that it is normally included in dictionaries alongside non metaphorical (basic) meanings” (Semino, 2008, p. 19). For example, in lines 25-26, Amos says: “(they) assigned, I took on the task of establishing a factory.” The Hebrew transliterated utterances hitilu, lakaxti lehakim mifal constitute conventional metaphors because their metaphorical meanings are included in the dictionary (Choueka, 1997).

28 In view of the fact that the distinction between deliberate and conventional metaphors is still not clear (Gibbs, 2011; Semino, 2008) and some researchers hold that “deliberate metaphors are not essentially different from other forms of metaphoric language” (Gibbs, 2011, p. 26), and in view of the fact that previous studies counting the density of metaphors in text were based on conventional and deliberate metaphors, the present study will focus on both.

29 Metaphors have been associated with the expression of emotion (Gibbs, 1994; Kövecses, 2003; Kupferberg, Green, & Gilat, 2008). Current studies associate linguistic metaphor density with the intensity of emotions that is expressed in the text (Cameron, 2009; Kupferberg, Green, & Gilat, 2008). For example, Cameron (2009) indicates that linguistic metaphor density ranges from 27 metaphors per 1000 words in classroom discourse to 100 words in emotionally-charged reconciliation talk. A current study (Kupferberg & Gilat, 2012) further shows that there were significantly more metaphors and metaphorical clusters in the stories of computer-mediated suicidal help seekers compared with non-suicidal help seekers.

30 Discourse-oriented students of metaphor have explored the functions of metaphors in discourse. Some metaphors constitute central organizing narrative devices (Kupferberg & Green, 1998, 2005) that often reveal implicit meanings (Kupferberg & Green, 2005; Semino, 2008). This metaphorical function is supported by diverse studies focusing on everyday conversation (Holt & Drew, 2005), troubled talk (Kupferberg & Green, 2005), reconciliation talk (Cameron, 2007), classroom discourse (Cameron, 2003), experts’ oral and written explanations (Cameron & Low, 2004), and political discourse (Mieder, 1997).

31 Metaphorical clusters (i.e., two or more metaphors occurring in a sequence and focusing on the same theme) are discursively realized as “sites of intensive work relating to the central discourse purpose” (Cameron & Stelma, 2004, p. 107). Clusters are produced at critical moments (Candlin, 1987, as cited in Cameron & Stelma, 2004) when interlocutors experience an external or internal obstacle undermining communication. For example, Amos uses a metaphorical cluster to conceptualize the meaning of the opposition that he was obliged to face when he was constructing the factory, 33-35).

32 Metaphorical clusters have been explored in face-to-face therapy (Pollio & Barlow, 1975), telephone and cyber troubled talk (Kupferberg & Green, 2008a, 2008b), reconciliation talk (Cameron, 2007; Cameron & Stelma, 2004), classroom discourse (Cameron, 2003), college lectures (Corts, 2006), experts’ explanations (Cameron & Low, 2004), Baptist sermons (Corts & Meyers, 2002), and business media discourse (Koller, 2003).

33 In the following example, a 25-year-old woman seeks help from a telephone hotline service because her parents allegedly abuse her mentally and physically. At a certain point in the interaction, the caller produces a cluster: “I swallow a lot. How much can I swallow? Are they making a doormat out of me? How much can I swallow? Do other people deserve to be treated like the kings of the world? Am I their doormat?” (Kupferberg & Green, 2005, p. 97).

34 This cluster is produced at a critical junction in the interaction when the caller is obliged to defend her credibility as a narrator. It is interesting to note that in this cluster metaphors work together (Kupferberg & Green, 2005) with repetition and rhetorical questions in the caller’s attempt to express her problem. Accordingly, to conceptualize her helplessness vis-à-vis other family members, the narrator repeatedly (and therefore emphatically, Buttny & Jensen, 1995) produces metaphors highlighting the positioning of “doormat” who is positioned by “the kings of the world.” In addition, the narrator strongly protests against the abusive family dynamics by inserting the metaphors in rhetorical questions (Quirk, Greenbaum, Leech, & Svartnik, 1985), which are syntactically interrogatives but semantically constitute a protest.

35 In sum, in order to explore the interactional dimensions of Amos’s narrative positioning, I espouse MPA that advocates the division of the story into levels so that I can distinguish between the interviewee’s interpersonal positioning in relation to the interviewer in the ongoing conversation (Level 1) as well as in relation to others in the narrated past (Level 2) and in possible future stories (Level 3).

36 I also adopt the idea underscored in the literature review that metaphors constitute organizing self-displaying linguistic tools that work with other language resources to reveal explicit and implicit dimensions of the narrator’s positioning at each of the levels. The following research question guides the analysis of Amos's life story: What are the characteristics of the interviewee’s narrative positioning as identified by MPA?

37 The methodology underlying MPA (Kupferberg, 2010b) is based on three assumptions. First, as I emphasized earlier, narrative time enables humans to shift from the level of the present (Level 1) to the levels of the past (Level 2) and the future (Level 3) and position themselves or be positioned by others at each level. Second, the analysis foregrounds the centrality of metaphorical language resources—the building blocks of the positioning levels working together with other language resources (Kupferberg & Green, 2005) and indicating where the interlocutors’ mental life is discursively located at each point in the story (Chafe, 1994).

38 Third, Level 1 is expressed via metaphors and other grammatical and lexical means that construct the interlocutors’ positioning in the ongoing conversation in relation to other interlocutors (see analysis of Level 1 units in the following section). Level 2, the level of past experience, is defined as what participants say about their life prior to the present (see our analysis of Level 2 units). Level 3, the future level, is defined as what participants say about their future plans and wishes. MPA also emphasizes the researcher’s construction of meaning at each of the three levels and the summary of this process at a fourth interpretive level that is presented in the findings and discussion sections of this article.

39 Methodically speaking, when I apply MPA, I divide the entire text into 3 levels (i.e., Level 1, Level 2 and Level 3) so as to extract meaning from each, and then construct an interface at a fourth level when the insights gleaned via microanalysis of the other levels are interpreted and associated with theory.

40 The hierarchically-ordered units of analysis comprise a story that is further divided into three level units (Ben-Peretz & Kupferberg, 2007; Kupferberg & Ben-Peretz, 2004). Level 1 units construct explicit or implicit interpersonal relations in the present ongoing interaction vis-à-vis the interviewer (e.g., Amos addresses the interviewer at the end of the story to elicit feedback from her, 59-60); Level 2 units construct past experience (e.g., Amos lists several positions that he held in the army, 13-15); and Level 3 units display the construction of a future level. The division of the text into Level units is justified by MPA's methodological tenets presented earlier in this section.

41 After the text is divided into Level units, single metaphors and metaphorical clusters are identified at each level. The level analysis and metaphor analysis of Amos’s story were conducted by two discourse analysts acquainted with MPA. Inter-rater reliability as tested by the percentage of agreement between them on unit allocation, metaphor identification, and function definition was 95%, 96%, and 94%, respectively. Cases of disagreement were discussed until an agreement was reached. Finally, the ratio between Level units 1-3 and the total number of units was computed in order to show which of level of positioning was more prevalent than the others. In addition, the density of metaphors was computed by the number of metaphors in the text per 1000 words. This step enabled me to compare metaphor density in Amos's life story with other studies (Cameron, 2009).

42 The examples presented in the article were translated from Hebrew by a professional bilingual translator who was careful to find adequate metaphors in English for the metaphors that were produced in the original text in Hebrew .

43 The division of the story into units via the allocation of Level units shows that Amos produced 127 Level units, 18% of which were Level 1 units, and 82% of which were Level 2 units. In addition, the density of metaphors in the text is 41 words per 1000. This density is higher than what Cameron (2009) defines as the lowest metaphor density (i.e., 20 metaphors per 1000 words) computed for classroom discourse. In Amos’s story, only 1 Level 3 unit was identified: “And…and…these days I go back and forth between thinking that I’m healthy and the future” (45-46).

44 In other words, descriptive statistics show that Amos chooses to focus on past experience (Level 2), but at times his mental life (Chafe, 1994) shifts to the present (Level 1) in order to handle interpersonal matters with the interviewer, or as a result of emotions that cause him to weep. It is evident that Amos shies away from the future and produces only few metaphors.

45 Amos uses Level 1 units to perform diverse functions in the interaction between him and the interviewer. First, the narrator uses various discursive markers to inform the interviewer where he makes a transition from one period of his life to the next. In this way, he also signals to her that he is capable of constructing a chronologically-ordered story that has a closure: “We at the first stage” (2), “Okay” (19), “That’s it” (35), “And that’s it” (59), “And that’s that. About myself. What else do you want to hear? Interesting?” (59-60). Using these Level 1 units, the narrator divides his life into three chronologically- ordered periods whose boundaries are clearly marked by him: Period 1—life until his release from the army (1-19); Period 2—life on the kibbutz until his stroke (19-35); and Period 3—life after his stroke (35-end). These chronologically-ordered periods should be distinguished from the Level units that indicate where Amos' s mental life is located at each point of the story.

46 Other Level 1 units express metalinguistic comments that the narrator produces in order to assist the interviewer in understanding the content of his narration: “a stroke” (35); “and I say as opposed” (37); “It bothers me quite a bit these days. Meaning the shift between disability and activity” (38-39); and “And that’s a long time. Very long” (48).

47 Finally, Amos uses Level 1 units to repair what he says (e.g., “I came— (they) brought me” (1-2). Self-repair and other-repair (e.g., when interlocutors repair themselves or are repaired by others, respectively [Schegloff, 2007]), frequently occur in unplanned speech (Ochs, 1979) when interlocutors attempt to express themselves.

48 The self-repair examples in Amos’s life story are listed below: “I came—(they) brought me” (1-2); “from the year…’42 …no…don’t remember, ’42” (10-11); “I was…and after that back to Gev” (24); “After that, (they) assigned me—(they) assigned, I took on the task of establishing a factory” (25- 26); “And it so happened that today the factory… When I established the factory it was…a bit of a problem in Gev” (29-30); “sometimes I…I think that I [suppressed weeping] am healthy today in (my) thinking (40-41); and “I was…when I was active” (50-51).

49 The self-repair work that is performed in lines 1-2 and 25-26 is particularly interesting because in these lines the narrator wavers between being an active agent and being under the care of others. There is one instance of other-repair when Amos’s wife assists her husband in recalling where they met as well as in positioning him as a person whose status was important (18-19). Viewing these self- and other-repair instances together, we see that they all indicate that the narrator's positioning is ambivalent.

50 Amos's wife’s intervention can be defined as an instance of co-narration that frequent in discourse (Jacoby & Ochs, 1995) and its contribution to sick people has been foregrounded in the literature. Of particular interest are the studies focusing on the co-construction of meaning when severe memory loss interferes, as in the case of Alzheimer patients (Freeman, 2011; Hamilton, 2008).

51 There are also two instances when the narrator starts weeping. It is plausible to assume that during the narration of Period 1, Amos’s emotions overwhelm him when he realizes how different he is at the time of the interview compared with the young and healthy man who was active in the underground army in the pre-state Jewish settlement (13). Then, in Period 3, the narrator weeps when he talks about being physically disabled but cognitively competent (40). In another Level 1 unit, the narrator actually verbalizes an online cognitive process when attempting to recall the past (11).

52 In conclusion, Level 1 non-metaphorical units perform various functions in the interaction. Using these units, the narrator positions himself vis-à-vis the interviewer as a competent conversation partner who complies with her request to unfold a life story, and who is capable of illuminating his narration via meta- linguistic coherence-inducing comments. I also interpret these units as revealing that at times the narrator's wavers between agency and helplessness in the face of old age and illness. It is noteworthy that there are only 2 metaphorical Level 1 units in lines 25-26: “(they) assigned, I took on the task of establishing a factory.” In Hebrew hitilu, lakaxti lehakim mifal is a conventional metaphor.

53 Level 2 metaphorical and non-metaphorical units show how Amos positions himself in the narrated past. Using non-metaphorical units, Amos produces a chronologically-ordered account of the events that took place in Period 1 (i.e., life until his release from the army, 1-19). A close examination of these units shows that Amos's positioning vacillates between positioning himself and being positioned by others. For example, Amos uses the first- person pronoun to position himself as an active agent (e.g., “I was…in the beginning a squad commander,” 13-14). At times, he is a recipient of other unknown people’s decisions (e.g., “Within this framework I was sent to the Palmach,” 8-9). The indefinite third-person “they” (Berman, 1979) is also present in the active voice (e.g., “before (they) had recruited all the Hachsharas,” 10). In other Level 2 units, he is a member of a collective “we” (e.g., “After that we moved over to the 4 ^th Battalion,” 12).

54 In lines 15-16, Amos produces an organizing story-internal metaphor (Kupferberg & Green, 1998, 2005) that summarizes the meaning of the military service in terms of “drifting.” The narrator uses the Hebrew verb hitgalgalti, literally, “I moved like a wheel.” The English translation does carry the mechanical movement but lacks the revolving nature of the movement.

55 This metaphor constitutes a succinct version of the more detailed description of Amos's experience (Kupferberg & Green, 1998, 2005), indicating that the narrator was positioned by others rather than being an active agent. In conclusion, metaphorical and non-metaphorical Level 2 units show that the narrator's positioning in Period 1 is ambivalent.

56 In Period 2 (19-35), Amos focuses on his life on the kibbutz. The narration of this period also unfolds like a historical account until the narrator attempts to express the meaning of the establishment of the factory (25-35). It is important to note that there are several self-repair Level 1 units that I discussed earlier (25-26, 27-28, 29-30) that slow down the flow of the narrative when the narrator attempts to conceptualize the meaning of the factory from his point of view.

57 Having described different roles that he performed on the kibbutz (20- 25), Amos tries to evoke the difficulties connected to the establishment the kibbutz factory. At this discursive junction, Amos produces a metaphorical cluster: “And in the beginning it limped along a bit. And then (they) actually began to run after me. Why did you create this white elephant and why that” (31-33).

58 The cluster enables the narrator to position himself via significant others in the past. In the first metaphor (“And in the beginning it limped along a bit”), the narrator expresses the meaning of the beginning of the construction process in terms of a disability. Then, in the second metaphorical Level 2 unit (“They began to run after me”), the narrator embeds the metaphor in a syntactic structure that positions him in a situation of helplessness (Bamberg, 1997c; Kupferberg & Green, 2005). Pronouns further emphasize that the collective indefinite “they” (Berman, 1979) is actually chasing the individual (“me”). The indefinite collective “they” repeatedly mentioned in the story in subject position—first in the family (1), then in the army (10) and the kibbutz contexts (25, 31, 32)—places the narrator in the position of the recipient of the action initiated by the collective “they.” This syntactic position often indicates helplessness and dependence. (See analyses in Bamberg, 1997b; Kupferberg & Green, 2005).

59 The third metaphorical Level 2 unit works together with constructed dialogue (“Why did you create this white elephant and why that”) and a rhetorical question. By using constructed dialogue (Georgakopoulou, 1997), Amos recycles the past voices of kibbutz members who were against the establishment of the factory. Following Bakhtin (1981), I interpret these voices as “double voices” that express what the others said as well as Amos’s anger and protest against these voices (Kupferberg & Green, 2005). The rhetorical question that is syntactically an interrogative, but semantically constitutes a protest (Quirk et al., 1985), further strengthens the narrator's emotion of anger that is expressed via constructed dialogue.

60 Why does the narrator produce this cluster at this point in his life story? It seems that the factory must have been the peak of the narrator’s career. Therefore, it was important for him to verbalize the difficulties in order to intensify the significance of his personal involvement and success. To accomplish this discursive task, the narrator uses metaphors integrated with other language resources. In this way, he manages to express what must have been quite difficult to express (Gibbs, 1994) by using non-metaphorical language. This interpretation is confirmed in lines 33-35 when the narrator produces an instance of metonymy (i.e., a figure of speech consisting of the use of the name of one thing for that of another with which it is associated [Merriam Webster Dictionary]). This instance of metonymy emphasizes that the factory is the only thing that supports the kibbutz in the present.

61 To conclude: in the second period of his life, the narrator positions himself as a man whose contribution to the kibbutz was valuable although it was not immediately acknowledged by others. In other words, Period 2 ends with a coherent statement of the narrator’s worth as a person who established a successful factory despite the difficulties.

62 In Period 3 (35-end), the narrator attempts to conceptualize the meaning of life after the stroke. Amos's narration lacks the historical account format that characterizes certain sections of the preceding periods. The narrator produces several single metaphors and a cluster when he reflects on this period.

63 Amos begins by conceptualizing the meaning of the stroke in terms of a resounding blow (Merriam Webster Dictionary) (“I got a zbeng,” 35). In the context of Amos’s life story, this is another organizing metaphor (Kupferberg & Green, 2005) that establishes an alternative division of Amos’s life into life before and after the stroke.

64 The narrator also conceptualizes other aspects of his daily life metaphorically. Using a cluster, he expresses the meaning of the consequences of the stroke in terms of a movement from the inside of the illness to the outside, resulting in keeping his cognitive ability intact: “I came out with an intact mind” (37-38). In Hebrew, the metaphor “intact mind” is rosh shalem, literally, “a head that is not broken and all its pieces are in the right place.” In addition, the narrator emphasizes the meaning of post-stroke life via other metaphors: “the shift between disability and activity” (38-39) and “So it took me out of the frame” (55-56). In Hebrew, “shift” is maavar— literally, “a passage”—and is metaphorical.

65 In Period 3, there are two additional language resources that enable the narrator to verbalize the meaning of life after the stroke. First, Amos produces generic, or habitual stories (Ochs, 1997) depicting his daily routine after the stroke (41-42, 49). Second, the narrator uses several negative Level 2 units that construct his inability (43, 44, 57-58). (See Kupferberg & Gilat, 2012, a study on the salience of negation in troubled talk).

66 We see, then, that Period 3 differs from the first two in its reflective style and the production of more metaphors when compared with Periods 1 and 2. It is evident that in this period, the narrator attempts to conceptualize the drastic change in his life that made him physically dependent on others. However, Amos repeatedly emphasizes that his ability to think has not been affected. The discursive movement between cognitive independence and physical dependence constitutes a central theme in Period 3.

67 In conclusion, analysis of Level 1 and Level 2 units and the absence of Level 3 units show that the narrator’s positioning is constructed via metaphors that often conspire with other language resources such as syntactic structure, pronouns, and constructed dialogue. In Period 1, the narrator is positioned by circumstances and unidentified others. In Periods 2 and 3, the narrator positions himself clearly, first as a useful kibbutz member who contributed by establishing a factory, and then in Period 3, by emphasizing that although he is positioned physically by the illness, cognitively he is well.

68 What are the characteristics of Amos’s positioning as identified by the metaphor-oriented MPA? The analysis shows that Amos’s mental life, as reflected by the life story that he unfolds (Chafe, 1994), positions itself in the past, and at times “travels” to the present. Amos avoids the future altogether. In addition, the analysis shows that the narrator produces only few metaphors. The metaphor-based interpretative process has foregrounded different positions, or voices (Hermans, 2008), that the narrator constructs in relation to others in the present and the past.

69 For example, Amos was an army officer who was compelled by the stormy events preceding and following the establishment of the State of Israel to drift “through the army” (15-16) without having any real control over the events. He was also an agentive kibbutz member who established a very successful factory despite enormous difficulties that threatened to undermine the construction process. A third position displays a chair-ridden dependent invalid lamenting his existence following the stroke, as well as a competent interlocutor whose “intact mind” is capable of communicating a relevant, well- designed story “seasoned” with meta-linguistic comments.

70 Why does the narrator “invest” in the narration of the past, frequently move to the present, but avoid the future? What is the contribution of the metaphors that Amos produces to the level analysis of positioning? Do the voices that he produces cohere, or hold together, within the life story? I shall first relate to Amos's “trans-level journeys” between the present and the past and his avoidance of the future.

71 One could argue that given the narrator’s age and physical condition, such discursive behavior is to be expected. However, I would like to propose an alternative interpretation based on the analysis of Level 1 and Level 2 units that was presented in the previous section. Amos’s “big story” (Freeman, 2006, 2011) is oriented first and foremost to the interviewer’s request to unfold his life story in the face-to-face interaction. This request determines Amos’s communicative goal in the interaction. To this end, he attempts to distance himself from his experience (Freeman, 2006) and construct a chronologically- ordered story. Therefore, he “invests” in the past.

72 However, this discursive agenda is undermined by various “obstacles” that are occasioned by the evolving demands of the interaction. The metaphor- oriented level analysis contributes by showing when and how Amos’s chronologically-ordered narration (Level 2) is interrupted, obliging him to shift to the present (Level 1). Accordingly, he intentionally draws the boundaries of the different periods in his life using Level 1 discourse markers, or “journeys” to the present, to assist the interviewer by “spicing” his story with meta- linguistic comments. In this way, the narrator positions himself vis-à-vis the interviewer as a competent conversation partner who is capable of relating to the requests of other interlocutors. Inadvertently, when he is overwhelmed by emotions, Amos’s story is interrupted by his weeping or by a comment relating to his momentary failure to retrieve an exact date from his long-term memory. In brief, I conclude that the narrator's positioning at one level or another is related to his discursive goal as well as to the contingent demands of the ongoing interaction. This claim is supported by three current studies (Kupferberg & Gilat, 2012; Kupferberg et al., 2012; Kupferberg & Hess, 2013) that were presented in the literature review.

73 The second question was: What is the contribution of the metaphors that Amos produces to the level analysis of positioning? Metaphor density in Amos's life story is low. Yet the single metaphors and metaphorical clusters that he does produce when his mental life “travels” to the past carry out important discursive tasks. They summarize central themes in Amos's life story (e.g., "a zbeng")and signal an extremely significant discursive junction where the narrator uses a cluster to position himself as a very valuable kibbutz member. In this way, story-internal metaphorical language foregrounds implicit meanings and enhances coherent narration.

74 Finally, do the various positions that the narrator produces cohere, or hold together, as dimensions of Amos’s self? To answer this question, I align myself with a current debate on the definition of narrative coherence (see evaluation and summary in Freeman, 2010) that challenges the dominance of the concept of coherence in the context of health communication because it marginalizes important phenomena identified in the discourse of the ill (Hyvärinen et al., 2010; Kupferberg et al., 2012, 2013) .

75 In this context, I argue that instead of evaluating Amos’s life story in dichotomous terms of coherence/incoherence that the definition of positioning presented earlier in the article adheres to, we should consider the story in terms of “degrees of positioning along a continuum ranging between full positioning and no positioning. This continuum should also be defined in terms of varying contextual factors that are relevant to each individual whose positioning is explored in research or examined in practice” (Kupferberg et al., 2012, p. 180).

76 Accordingly, we should define Amos’s strengths bearing in mind his age and physical limitations. In so doing, it becomes evident that Amos is a man who acknowledges his physical limitations, attempts to overcome his emotions, and succeeds in finding solace in his ability to think, narrate a well- designed story, and reflect on past successes that attest to how significant he was.

77 What are the strengths and limitations of MPA presented in this article? The contribution of the method to narrative analysis can be summarized in theoretical, methodological, and practical terms. First, the method is based on a theoretically justified procedure that foregrounds the centrality of the narrator- oriented levels and their re-assembling and interpretation at a fourth level of analysis. Another theoretical contribution of the method is its reliance on story- internal metaphors and other language resources, which constitute the building blocks on which interpretation is based.

78 Methodologically speaking, the study foregrounds the contribution of the level analysis that allowed us to read Amos's text and be attuned to changes in his positioning. In addition, in the absence of para-linguistic cues that a recording of the life story could have provided, the reliance of MPA on self- displaying metaphors as well as other language resources strengthens the interpretive process by grounding it in the empirical evidence produced by the narrator himself. In this way, the analysis goes beyond mere content-analytic coding that often “locks aspects of the interaction” (Maynard & Heritage, 2005, p. 428) into a set of categories.

79 From a practical point of view, the study shows that the narrative discourse of the ill should be explored attentively even in situations when humans are unable to fully voice who they are (Kupferberg et al., 2012). Therefore, it is important to equip practitioners in the institutional settings with the appropriate tools and training that will enable them to be attuned to the discursive strengths and weaknesses of the other interlocutors in a given context.

80 What are the limitations of the analysis? First, it is quite difficult to analyze a face-to-face interaction in the absence of paralinguistic cues (Kupferberg, 2008). Accordingly, the analysis heavily relies on language resources that require thorough acquaintance with theoretical and methodological discourse analytic frameworks.

81 In addition, the level analysis of positioning relies on native speakers’ use of varying language resources including metaphors. In the case of non- native speakers, the researchers’ work of identification, description, and interpretation might be severely undermined.