Resistance to Metaphor in Dementia Discourse –
Plants, Zombies, and Brain Training
University of Amsterdam | Master’s Thesis
Student E.M. Turkenburg
Number 11257717
Supervisor Prof. Dr G.J. Steen
Second readers Dr G. Frezza
Dr R. Pilgram
Research Master Communication and Information Studies
Rhetoric, Argumentation Theory and Philosophy
Table of contents
Abstract ... 4
1. Introduction ... 1
2. Theoretical framework ... 3
2.1. Dementia perception in society ... 3
2.2. Metaphor and figurative framing ... 5
2.3. Resistance to persuasion ... 9 3. Research question ... 11 4. Method ... 13 4.1. Ethics statement ... 13 4.2. Design ... 13 4.3. Participants ... 13 4.4. Materials ... 14 4.5. Measures ... 15 4.6. Procedure ... 16 4.7. Analysis ... 17 5. Results ... 18 5.1. Narrative ... 18 5.2. Frame ... 19 5.3. Narrative x Frame ... 19
5.4. Moderator variables – Attitude, Perceived knowledge, Prior knowledge ... 21
6. Discussion and conclusion ... 23
6.1. Narrative and Frame ... 23
6.2. Attitude ... 24
6.3. Perceived and prior knowledge ... 25
6.4. Directions and limitations ... 25
7. Acknowledgements ... 27 8. References ... 28 9. Appendices ... 32 9.1. Prior knowledge-questions... 32 9.2. Attitude-questions ... 32 9.3. Vignettes ... 33
Abstract
This master’s thesis presents research into the resistance evoked by metaphorical and non-metaphorical frames in dementia discourse. The effects of these framing types on resistance are studied in relation to the two opposing narratives of plasticity (positive) and decline (negative), which dominate dementia public discourse (Frezza, In press). These narratives are guided by metaphor, a figurative framing type which is arguably a powerful persuasive tool to influence reasoning (e.g., Burgers, Konijn & Steen, 2016; Thibodeau & Boroditsky, 2015; 2017). Little is known about resistance to such figurative frames, but it is evident that persuasion potentially evokes resistance (e.g., Zuwerink Jacks & Cameron, 2003). The aim of this research was to bring the field of metaphor studies and the field of resistance research closer together. To test the effects of metaphoricity of frame and text narrative on amount and type of resistance evoked, a 2(metaphorical/non-metaphorical frame) x 2 (plasticity/decline narrative) multiple-message experiment was conducted. Effects of existing attitude, perceived knowledge, and prior knowledge were taken into account. Results indicate a gap between prevalent narrative in discourse and prevalent attitude displayed by people. Important factors to take into consideration include emotional proximity of discourse theme and risks of socially acceptable answers. Further research is needed to advance the marriage of metaphor (framing) research and resistance research and to provide more insights into the mechanisms constituting dementia perception in society.
1. Introduction
Three metaphors are used in this statement about the importance of language: language is a foundation, language is a glue, and language is a weapon. This example from the sci-fi movie Arrival – which is more about language than about aliens and a true must-see – is a good example of how ubiquitous metaphor is in our language and how it can be used to frame.
To discuss social and political issues, we rely heavily on metaphor (Thibodeau & Boroditsky, 2013). Metaphors pervade discussions of critical issues, making up as much as 10-20% of natural discourse (Steen, Dorst, Herrmann, Kaal, Krennmayr & Pasma, 2010). Unlike in former times, when metaphor was just seen as an ornamental language device, today we know that metaphors matter. There is even research which shows signs of metaphor affecting people’s reasoning. This would mean that metaphor can be used as a powerful tool of persuasion.
We still need to know more about the conditions under which metaphorical persuasion takes place. Research on the differences between metaphorical and literal language and the boundary conditions of metaphorical persuasion is inconclusive (e.g., Reijnierse, Burgers, Krennmayr & Steen, 2015). A factor which has so far not been considered in research on metaphor’s persuasive power is the possibility of evoked resistance. Persuasive tools have the potential to take away resistance, but a persuasive tool also brings along the risk of evoking even more resistance (e.g., Fransen, Smit & Verlegh, 2015).
One field of discourse where the persuasive application of metaphor would be particularly welcome is the field of dementia. More and more people state that there is a need for a more social approach to dementia (The, Pols & Pool, 2017), in which dementia patients are not seen as becoming ‘plants’ or ‘zombies’, but are treated with respect as active individuals, who can train their brains. Recent research has shown that dementia discourse is highly metaphorical and ruled by two encompassing narratives for the brain: a positive plasticity narrative and a more negative decline narrative (Frezza, In press).
The present research looks at the resistance evoked by metaphor in dementia discourse, in relation to two opposing narratives of plasticity and decline. We examine the amount and different
Language is the foundation of civilization. It is the glue that holds a people together.
It is the first weapon drawn in a conflict.
types of resistance evoked by texts displaying either a decline narrative or a plasticity narrative. The texts are designed to support the narratives with either a metaphorical or a non-metaphorical frame. In doing so, we will also take into account the impact of prior attitude towards dementia, since a more decline-oriented attitude is expected to result in more resistance towards a plasticity-based text, and a more plasticity-oriented attitude is expected to result in more resistance towards a decline-based text. Based on suggestions in earlier research (e.g., Zuwerink Jacks & Cameron, 2003) we also take into account the role of prior knowledge and perceived knowledge of the topic of dementia.
Before elaborating on the experimental method employed in this study, I will provide a more encompassing background on dementia, metaphorical framing, and resistance. Then, I will present the research question and corresponding hypotheses. After explicating the method employed in this study, I will present the results, followed by a discussion and conclusion based on these results.
2. Theoretical framework
2.1.
Dementia perception in society
Approximately 50 million people worldwide live with dementia and many people know a family member or close friend living with this condition (WHO, 2017). Dementia is now the seventh leading cause of death in the world and the number of dementia patients is expected to increase to 82 million in 2030 (WHO, 2017). It is no wonder dementia is receiving more and more attention and acknowledgement in public discourse. Former British prime minister David Cameron described dementia as a ‘national crisis’ (BBC, 2012), newspapers speak of an ‘Alzheimer’s epidemic’ (e.g., Haseltine, 2018), and the World Health Organization has recognized dementia as a public health priority and calls for more awareness (WHO, 2012).
The condition is a ‘hot topic’, so it is surprising that there are still many misconceptions and incorrect images about dementia and the people suffering from it. Research by Alzheimer’s Research UK (2015) indicates for example that 75% of people are unaware that they can affect their own risk of developing dementia. Still, for 39% of people under 60 years and for 52% of people of 60 years and above Alzheimer’s is the disease they are most concerned about. After getting diagnosed with dementia, 85% of people want to stay at home as long as possible, but less than half of the general public believes that this is possible (Alzheimer’s Research UK, 2014).
Platforms and organizations from all over the world call for a ‘dementia-friendly society’. This seems highly necessary. People with dementia easily lose their social networks (Sweeting & Gilhooly, 1997) and nearly 70% of people feel isolated following a dementia diagnosis (Alzheimer’s Research UK, 2016). Dementia patients suffer from a poor self-image, have greater difficulty in accessing support services, and may experience discrimination (Naue & Kroll, 2008). According to Alzheimer’s Research UK (2016) over 40% of the general public believes that a dementia patient does not benefit from spending time with loved ones he or she no longer recognises.
A recent episode of the Dutch news show ‘Nieuwsuur’ (NTR - NOS, 2018), showed how most people view dementia patients as sad, old people wasting away in nursing homes, not remembering their children’s names. People with dementia suffer from this image, which is focused on the very last stage of dementia. They become lonely since others see them as ‘sick’ and ‘crazy’. Friends and family visit less since they think it’s ‘too confronting’, or they want to remember ‘the old version of the person’. All the while the patients themselves are still the same and do not feel like they are being taken seriously.
The negative image of dementia in society seems to be reinforced by the media (Van Gorp & Vercruysse, 2012). Of people with dementia 55% are in the mild stage, 32% are in the moderate stage and 13% are in the severe stage (Prince et al., 2014). Studies on the image of dementia show how the media mainly focus on this last, severe stage of the condition as representative of the entire process. Besides this, dementia patients rarely speak for themselves in the media and the emphasis usually lies on how heavy the illness is for family and caregivers of the patient (Van Gorp & Vercruysse, 2012). Dementia is often framed in a negative way in public discourse and there is a lack of alternative counter-frames to adjust the public perception of dementia (Van Gorp & Vercruysse, 2012).
The dominance of this negative dementia image is confirmed by Frezza (in press). Frezza’s research shows how in dementia public discourse, two encompassing narratives for the brain can be distinguished. There is a more negative ‘decline-narrative’, based on the traditional account of dementia as a downward trajectory. The brain is decaying and can only go downhill from here, people are depicted as ‘zombies’, leading a ‘vegetable existence’. This negative view leads to hopeless associations of passiveness. This narrative is most commonly used and conventional in public discourse. A more hopeful narrative is that of ‘brain plasticity’, which depicts the brain as a material organ, something resilient which can be trained and the condition of which can be improved.
The decline narrative seems to call for a more passive treatment of comfort and softening the consequences because the situation cannot be improved, just palliated. The innovative, more positive plasticity-narrative points to a more active role of the individual. It is not used as often but also fits the current medical knowledge of the illness, namely that one’s condition can be improved by sports, music, and activities in general (Alzheimer’s Research UK, 2018). Dominance of either one of these narratives could have an impact on the way we treat dementia patients. It would, therefore, be useful to find out how the use of these narratives in public discourse could be steered.
A potential tool to influence the image of dementia in society is the use of metaphorical frames. Besides finding the two narratives for the brain, Frezza (in press) also showed how these narratives are usually accompanied by metaphor on both the individual level (e.g., ‘someone turning into a zombie’) and the societal level (e.g., dementia being a ‘tidal wave, threatening our society’). Metaphors provide frames of thinking about societal topics, so the ways metaphors change can reveal how conceptualizations of societal topics change over time (Burgers, 2016). Studying shifts of metaphors can be used to study changes in social discourse, but perhaps metaphor can also be used to actively make that change and influence the conceptualization of societal topics.
2.2.
Metaphor and figurative framing
Metaphor is ubiquitous in all language use, both spoken and written (e.g., Cameron, 2008). The persuasive power of figurative language like metaphor has been acknowledged since the days of antiquity (Burgers et al., 2018). For a long time, however, the persuasive power of figurative language was merely attributed to its ‘artful deviation’ (McQuarrie & Mick, 2003). Metaphor was seen as a purely ornamental stylistic device on the surface of language, used to vivify our speech and writing: nothing more, nothing less. Such a stylistic device can alter a text to be more appealing and interesting, but was it in no way supposed change anything about the message content on a conceptual level (e.g., Burgers, Konijn & Steen, 2016).
This perception of metaphor changed in the 1980s when renewed interest in metaphor research caused what is often called ‘the cognitive turn’ (e.g., Landau, Robinson & Meier, 2014). Research shifted from perceiving metaphor as a matter of just language to seeing it as a matter of thought (e.g., Landau et al., 2014; Burgers et al., 2016). In their Conceptual Metaphor Theory (CMT), Lakoff and Johnson (1980) argue how metaphors are cross-domain mappings, involving cognitive concepts in the mind (source domain), that are used to understand more abstract concepts (target domain). Since the birth of CMT, metaphor is not just seen as a matter of words anymore; it is fundamental to human thought.
Metaphor is often claimed to have the ability to influence reasoning through hiding and highlighting information (Reijnierse, Burgers, Krennmayr & Steen, 2015). By accessing metaphorical source domains, we can be led to pay attention to specific aspects of a target domain in such a way that other aspects of the target are down-played or even left aside. Because of this ability of metaphor to influence reasoning, it is argued by Steen (e.g., 2008; 2011; 2017) that we should not only study metaphor in language (linguistic dimension) and in thought (conceptual dimension). Instead, Steen advocates the addition of a third dimension in which metaphor should be studied: communication.
Where metaphor in thought relates to metaphors’ conceptual function to offer frameworks for concepts, the communicative function of metaphor is to change people’s perspective on a referent or a topic. Studying metaphor on the communicative dimension is partly about whether metaphor is used deliberately or non-deliberately; whether there is the intention to change the addressee’s perspective. Steen (e.g., 2008) calls the use of metaphor on the language level ‘naming’, the conceptual function of metaphor in thought ‘framing’, and the communicative function ‘changing’ (perspective).
An encompassing theory of using figurative language (like metaphor) to frame any specific referent or topic was presented by Burgers et al. (2016). They present the theoretical perspective of figurative framing, to explain framing across a variety of societal issues. Frames are typically said to consist of two devices. There is a framing device, which is the surface-component, the linguistic packaging of the frame; and there is a reasoning device, comprising the information given in a text, the
conceptual level. Burgers et al. (2016) state that key figurative language types, amongst which metaphor, work as both framing devices and reasoning devices. The traditional view of figurative language as a mere framing device is challenged, by stating that figurative language can fulfil both functions of framing and reasoning. The prediction is made that figurative frames can be used to attack or maintain existing frames.
Burgers et al. (2016) state that figurative framing is an important concept in communication, but they only study figurative language on the language-level (as a framing device) and on the thought-level (as a reasoning device). As stated before, Steen (e.g., 2008; 2011; 2017) approached metaphors on three levels: naming (language), framing (thought), and changing (communication). Krippendorff (2017) however, proposes to redefine framing as an act of communication, concentrating on communicative behaviours (like the use of linguistic tropes such as metaphors). The current research proposes to combine the classifications by Steen (e.g., 2008), Burgers et al. (2016), and Krippendorff (2017) and consider metaphor in language as naming, metaphor in thought as reasoning, and metaphor in communication as framing.
Robins and Mayer (2000) investigated whether and when metaphor influences people’s decisions about how to resolve dilemmas in everyday domains. In six experiments they found evidence for the metaphor framing hypothesis: the idea that metaphorical frames affect reasoners’ judgements about text-based dilemma’s. Based on the metaphor framing hypothesis, Robins and Mayer predicted that reasoners would be more likely to agree with dilemma resolutions that are consistent with the metaphorical frame and that they generally will not indicate awareness of the metaphors’ influence on their judgements. In three of the experiments, evidence was also found for the metaphor processing termination hypothesis: if a reasoner possesses other knowledge to help in making sense of a text description or a dilemma, the reasoner is less likely to use the metaphor, consequently weakening the metaphor framing effect.
The (covert) impact of metaphorical frames on reasoning has also been studied by Thibodeau and Boroditsky (2011; 2013) in their (famous) crime-experiments. They explored the role of metaphor in reasoning about a domain of societal importance. In their first study, Thibodeau and Boroditsky (2011) framed crime in a city as either a beast or a virus. Participants were presented with a short text containing one of these two metaphors, which were explicitly stated. Subsequently, they were asked to write down their solution for the crime problem. Participants generally tended to choose more frame-fitting solutions (more forceful and enforcement-oriented for the beast frame and a more social reform-oriented approach for the virus frame). In a follow-up study in 2013, Thibodeau and Boroditsky asked participants to choose the best solution from a selection of possibilities, instead of asking to come up with their own solutions. Even with this set of options to consider and compare, people choose a more frame-consistent solution.
A critical review of Thibodeau and Boroditsky’s work is presented by Steen, Reijnierse and Burgers (2013), who claim that the results reported in their 2011 and 2013 papers leave room for alternative explanations. In a modified replication of Thibodeau and Boroditsky’s studies, Steen et al. (2013) added a non-metaphorical control-condition and a pre-exposure measure of political preference. They found a general shift towards more enforcement-oriented solutions as a solution for crime, regardless of the metaphorical frame (virus or beast). People generally seemed to favour a strong, enforcement-oriented response to crime as if it were a beast. Steen et al. (2013) state that this suggests that there may be certain boundary conditions under which metaphors can affect reasoning in different ways. These findings match Robins and Mayer’s (2000) confirmation of the metaphor processing termination hypothesis based on existing prior knowledge.
Also following up on Thibodeau and Boroditsky’s work, Reijnierse et al. (2015) examined the influence of the extension of metaphor across the text. Reijnierse et al. (2015) hypothesized that extending a metaphorical frame in a text by increasing the number of metaphorical sentences will lead participants to display higher ratings of the perceived effectiveness of policy measures that are in line with that frame. The results pointed to an effect in the opposite direction of the hypothesis: extendedness facilitated metaphor rejection. These results suggest that metaphorical framing effects may be subtler than assumed. Reijnierse et al. (2015) also suggest that the theme of the experiment may have been of influence on this outcome. People may have favoured the enforcement-oriented solutions because crime remains an issue that requires immediate action, even if it is described as a long-term problem.
Thibodeau and Boroditsky (2015) further explored a metaphor framing effect in the crime context. Their results reveal a shift in the overall landscape of opinion on the topic of crime, but also show that metaphors continue to have an influence on people’s reasoning about crime. One important conclusion Thibodeau and Boroditsky (2015) draw, is that it can be difficult to study framing in a dynamic real-world context.
In their overview article on the way linguistic metaphor scaffolds reasoning, Thibodeau, Hendricks and Boroditsky (2017) express the need for a clearer image of boundary conditions for the effects of metaphor. They state that linguistic metaphor can guide thought and behaviour and they formulate a number of pressing questions for the field of metaphor research, amongst which the need for more information on conditions under which metaphor is most influential.
Two potentially influential elements on metaphorical reasoning mentioned by Thibodeau et al. (2017) are prior knowledge and attitude. They argue that prior knowledge about a subject affects the experience of aptness of the metaphor, and aptness of metaphor encourages more congruent inferences. Deep-seated attitudes make people less vulnerable to impact by a metaphor. It is, however, unclear how much prior knowledge and how strong an attitude is necessary for influencing thinking.
The influence of prior knowledge has also been proposed by Burgers et al. (2016), who state the prediction that when presented with a figurative frame of an unknown subject, participants with low knowledge will be affected more. In such a case, the frame makes the unknown issue more comprehensible by providing additional information.
Thibodeau et al. (2017) also raise questions about the advantage of distinguishing between literal and figurative language in research on the persuasive power of metaphor and ask whether there really is a difference between literal and figurative language. Many studies that examine the persuasive power of metaphor do not take non-metaphorical frames into account (Robins & Mayer, 2000; Thibodeau & Boroditsky 2013; 2015). Indeed, Thibodeau and Boroditsky (2015) have explicitly argued against the use of non-metaphorical frame, since they differ from metaphorical frames on multiple linguistic dimensions (e.g., valence, vividness, conventionality). Comparison would therefore be meaningless.
By contrast, Reijnierse et al (2015) argue that a non-metaphorical condition is necessary to determine whether findings can be attributed to the metaphoricity of the frame Without a non-metaphorical condition, there is no way of telling whether the effects of a frame should be attributed to the metaphoricity of that frame, or to the frame in general. Hartman (2012) also points out that many scholars often do not distinguish the metaphoricity from the frame, which makes it unclear whether it is the metaphoricity that influences reasoning, or whether it is the effect of framing in general. Furthermore, Brugman, Burgers and Steen (2017) assert that framing research from a linguistic perspective should not exclusively focus on metaphorical frames when studying the effects of metaphorical frames. Their study on research regarding political framing experiments demonstrates that metaphorical frames can be found in a substantial number of cases (even in political framing experiments that do not explicitly deal with metaphor). Two in three experiments, however, did not contain metaphorical framing, and five in six frames were non-metaphorical.
In sum, empirical studies investigating the effects of metaphorical framing show mixed results. Some research shows great influence of metaphorical frames on reasoning (e.g., Thibodeau & Boroditsky 2013; 2015). Other studies report no effects of frame-metaphoricity whatsoever (e.g., Steen et al., 2014; Reijnierse et al., 2015). Many scholars emphasize the influence of boundary conditions, such as attitude and prior knowledge.
One potential boundary condition which has been thoroughly neglected, however, is the evocation or reduction of resistance by metaphor. The field of metaphor research has heretofore mainly focussed on what makes metaphor more or less persuasive. The resistance language users show to metaphor’s persuasive power has not been considered. Research on resistance to persuasion has, on the other hand, not paid much attention to the difference between resistance evoked by literal and figurative language. It is the aim of this study to bring these two issues together.
2.3.
Resistance to persuasion
Despite all the research on persuasive tools like figurative framing, persuasive attempts often have limited impact because people are motivated to resist persuasion (Ringold, 2002). Resistance can be seen as the key to persuasion: if there were no resistance, we would have no need to persuade people. Still, persuaders tend to only focus on the persuasiveness of their message and do not acknowledge that persuasion and resistance go hand in hand.
When uttering a persuasive message and encountering resistance, a sender usually chooses to counter this resistance by using a so-called ‘alpha strategy’ (Knowles & Linn 2004). Alpha strategies deal with resistance by overwhelming it. When a buyer is in doubt, the alpha approach is to overpower the resistance with rhetoric (describing the benefits) or incentives (adding benefits). The problem is that alpha strategies do not diminish the concerns and apprehensions that a customer had, even when they are successful in persuading (Knowles & Linn 2004). Another, less used strategy is to directly target the resistance, the ‘omega strategy’. Instead of strengthening the persuasiveness of the message, the resistance is weakened by finding ways to avert, remove, or redirect resistance. With resistance reduced or removed, the customer is free to pursue the features and benefits that attracted them in the first place. You don’t need to add any incentives or explain the benefits any differently (Knowles & Linn 2004).
Decreasing resistance can be desirable for more commercial goals, such as a fast-food restaurant wanting people to buy their newest burger, but also for non-profit goals, like in a government campaign against smoking. Besides decreasing resistance, one might also want to increase resistance in some cases, to arm people against persuasive attacks, like when a health-organization want people to form more resistance against persuasive attempts from the fast-food restaurant mentioned earlier. To be able to use and steer resistance, we need to know more about how and why it is evoked and what form(s) it can take.
The counter strategies people adopt to resist persuasion and the motives for using specific strategies were brought together in a framework by Fransen, Smit and Verlegh (2015). One important motive for resistance is the ‘concern of deception’, which is comparable to what is often called the forewarning effect (Lee, 2010; Petty & Cacioppo, 1977; 1979). People may feel like or think that someone is misleading them and could, therefore, start to resist the persuasive attempt. This motive is activated when persuasive tools are recognized, which can happen both consciously and unconsciously. How prone someone is to show this type of motivation depends on several personal characteristics (Fransen et al. 2015; Lee, 2010; Nelson, 2016).
Earlier research on different resistance strategies evoked by persuasion was conducted by Zuwerink Jacks and Cameron (2003). To examine the variety of strategies used to resist persuasion, participants were asked to identify an attitude or belief that was highly personally important to them
and which they were not likely to change. They were then asked to write an essay describing how they would respond when someone or something would challenge that attitude of belief. The essays were analysed with an a priori coding scheme in which six expected strategies were listed: attitude bolstering (generating supporting thoughts for the own, original attitude); negative affect (having a negative emotional response); selective exposure (ignoring the message by leaving the situation or ‘tuning out’); counterarguing (rebuttal of the message arguments); source derogation (rejecting the source’s validity); social validation (bringing to mind others who share the original, own attitude). In the decoding process one unexpected, seventh strategy was added: asserting confidence (asserting that nothing or no one can change the original opinion). The data showed that attitude bolstering was the most prevalent strategy. It was found in 50 percent of the essay responses.
In two follow-up studies, the likelihood of each found strategy being used to resist persuasion was assessed. Participants were first asked to list their attitude towards abortion (study 2) or the death penalty (study 3) and indicate how personally important this topic was to them. After that, the participants were presented with seven responses people could have to a persuasive attack (each response representing one of the strategies found in the first study) and rated the likelihood of having each response. Attitude bolstering and counterarguing appeared to be most likely to be used for resisting persuasion, whilst source derogation, negative affect and selective exposure were rated least likely. The different subjects (death penalty and abortion) did not lead to different results.
In a final study, Zuwerink Jacks and Cameron (2003) investigated the actual use of the most likely strategies. Firstly, participants were asked to list their attitude towards and the personal importance of the topic of death penalty. They then listened to a speech against the death penalty, followed by a thought-listing task in which the participants were asked to give their final attitude towards death penalty, their affective response and an evaluation of the source. As expected from the previous studies, attitude bolstering and counterarguing were the most prevalent responses to the persuasive message.
3. Research question
As stated earlier, the narratives in dementia discourse are guided by metaphor. If we can use metaphor to frame, we might be able to steer the image of dementia in society by using metaphor. It would, therefore, be useful if we could establish how resistance is evoked by metaphors in dementia discourse.
Rhetorical figures, like metaphor, are said to increase message elaboration (e.g., Hoeken, Swanepoel, Saal & Jansen, 2009) and positively influence persuasiveness of advertisements (Van Mulken, LePair & Forceville, 2010). However, when these figures are recognized as persuasive tools, they might actually evoke resistance instead of persuade. So far, studies on resistance to persuasion have not considered the potential influence of figurative language as a persuasive tool. On the other hand, studies on the persuasive effects of metaphor have not considered the potential resistance evoked by this framing- and reasoning device.
The goal of this research is hence to contribute to knowledge on how to effectively use metaphor to steer resistance in communication, focusing on the case of dementia perception in society. The research question to be answered in this research is:
RQ. What kind of resistance is evoked by metaphorical decline/plasticity frames dealing with the topic
of dementia and how does this differ from the amount and types of resistance most evoked by literal decline/plasticity frames dealing with dementia, depending on prior attitude towards the topic and perceived and prior knowledge about the topic?
As noted in the previous chapter, scholars are not in agreement about the inclusion of literal language-conditions in research on metaphor’s persuasive power. However, research on resistance to persuasion has so far not considered figurative – let alone metaphorical – language use. This is why in this research, metaphorical language and literal language will be compared to investigate the following two opposing hypotheses:
H1a. Metaphorical frames do not differ from literal frames when it comes to the amount and
types of resistance they evoke.
H1b. Metaphorical frames differ from literal frames when it comes to the amount and
We know that two narratives for dementia are prevalent in dementia discourse: decline and plasticity. These narratives depict the attitudes towards the illness of dementia that are present in society. The factor of participants’ initial attitude is included in our research question because people tend to resist something more when it clashes with their initial attitude towards a topic (Zuwerink Jacks & Cameron, 2003). We hypothesize the following:
H2a. Participants retaining a decline-oriented attitude towards dementia show more resistance
towards texts in which a plasticity narrative is propagated, regardless of metaphoricity of the frame.
H2b. Participants retaining a plasticity-oriented attitude towards dementia show more
resistance towards texts in which a decline narrative is propagated, regardless of metaphoricity of the frame.
In their research, Zuwerink Jacks and Cameron (2003) showed how perceived knowledge of a subject marginally predicted use of the counterarguing strategy and significantly predicts the assertion of confidence strategy and selective exposure strategy. This suggests that when people think they know a lot about something, they are confident and sure about their opinions, so much that they are willing to assert that confidence or even walk away and not listen to opposing views. This research will also look into the effect of perceived knowledge on resistance. Contrary to Zuwerink Jacks and Cameron, who only looked at literal language, the influence of perceived knowledge is measured for both literal and metaphorical language.
There are different types of resistance, evoked under different conditions, similar to the boundary conditions for metaphor persuasion. Zuwerink Jacks and Cameron (2003) only measured perceived knowledge of participants but did not take into account their actual existing prior knowledge. Several scholars (e.g., Thibodeau & Boroditsky, 2017; Burgers et al., 2016) mention the potential impact of prior knowledge as a boundary condition for the effects of metaphor. To account for the potential influence of perceived and prior knowledge, the last hypotheses to be tested in this study is:
H3: The resistance evoked by metaphorically and non-metaphorically framed texts advocating a
decline or a plasticity narrative for dementia is moderated by (a) perceived knowledge, and (b) prior knowledge.
13
4. Method
To answer the research question, an experiment was conducted investigating the impact of metaphorical and non-metaphorical frames, employing decline- and plasticity narratives on the amount and types of evoked resistance.
4.1.
Ethics statement
Data were collected in accordance with ethical guidelines of the Amsterdam Institute for Humanities Research (AIHR). The study was approved by the Ethics Committee of the Faculty of Humanities of the University of Amsterdam (UvA). Participants took part in the study based on free-will and received a remuneration of 1.25 British pounds for their participation.
After participants read an information sheet on the first part of the survey, their informed consent was recorded by self-chosen continuation with the study. The information sheet included the guarantee that all data would be processed anonymously and that participants could quit the study at any time without having to provide a reason.1
4.2.
Design
A study was conducted employing a 2 (frame: metaphorically framed/non-metaphorically framed) x 2 (narrative: plasticity /decline) x 4 (text list) mixed design. Both frame or and narrative were within-subject variables. Four text lists were used to increase external validity. Every participant was appointed to one of the four text lists, therefore ‘list’ served as between-subject factor. Attitude, perceived knowledge, and prior knowledge were treated as covariates.
4.3.
Participants
Data for this experiment was collected in June 2018. Participants were recruited through the website Prolific Academic2 and were compensated with 1.25 British pounds. Before data were collected, sampling criteria were set. Participants were only deemed eligible to take part in this research if English was their native language, they were adult (18 or older) residents of the United Kingdom, and if they were not experiencing any (mild) cognitive impairment or dementia.
We expected participants to spend 10 minutes on taking the survey. Participants were allowed to spend a maximum of 45 minutes for filling out the survey. The average time spent was 7 minutes
1 The full text of both the Information Sheet and the Informed Consent form can be found in the Appendix,
section 9.4.
and 28 seconds, SD=3.95, Max.=28.98, Min.=1.67. A total of 140 participants completed the online survey, every participant saw every condition. There were 34 participants in text list A, 33 in text list B, 38 in text list C, and 35 in text list D. Age was not recorded, but eligibility to take part in this research required participants to be 18 years or older.
The majority of the participants (72,1%) was female, 25.7% was male and the remaining 2,1% did not want to specify their gender or listed ‘other’. Of our participants, 52.1% received higher education, 31.4% received further education, 14.3% received secondary education, and 2.1% stated ‘other’ when asked about the highest level of education they received. The distribution of both gender (X2(9)= 6,97, p=.640) and education (X2(9)=4.98, p=.836) was equal across the four different text lists.
4.4.
Materials
Four vignettes (short texts) were designed, inspired by the vignettes used by Robins and Mayer (2000) and Thibodeau and Boroditsky (2011; 2013). In these texts, participants could read a statement about people with dementia, and the way to deal with the illness and its consequences. One text read about dementia patients in general, one text was about dementia patient Ms Jones, in one text someone talked about their grandmother, and the last text was someone talking about their father who suffers from dementia.
All four texts were manipulated in such a way that there were four conditions for each text (sixteen texts in total). Every condition presented a mix of either a metaphorical frame or a non-metaphorical frame and either a decline or a plasticity narrative. The four conditions were labelled DM (Decline x Metaphor), DN (Decline x Non-metaphor), PM (Plasticity x Metaphor) and PN (Plasticity x Non-metaphor). From these sixteen texts, four lists were created. Each list contained one version of every text and one text in every condition. The distribution of texts and conditions over the four lists is presented in figure 1.3
The average length of all texts was 72 words (SD = 10.72). Average text length did not differentiate between the four lists (F(3,12)=.099, p=.959), nor did it differ between condition (F(3,12)=.055,
p=.982).
3 All sixteen texts can be found in the Appendix, section 9.3.
Text 1 Text 2 Text 3 Text 4
List A DM PM DN PN
List B PM DN PN DM
List C DN PN DM PM
List D PN DM PM DN
4.5.
Measures
A questionnaire was designed to measure the seven types of resistance evoked by the texts, as well as attitude towards dementia, perceived knowledge about dementia, and prior knowledge about dementia.
Resistance was divided into items for each of the seven types of resistance found by Zuwerink Jacks and Cameron (2003), which were combined to create the calculated variable of Total Resistance. The different strategies are operationalized as potential reactions to the texts, similar to what Zuwerink Jacks and Cameron (2003) did. To determine scores for the several types of resistance, participants score the likelihood of giving each of the responses to the vignette they just read on a 1 (very unlikely) to 7 (very likely) scale. To describe each strategy, Zuwerink Jacks and Cameron (2003) used two elaborate and complicated sentences. For the current research, only one sentence per strategy was used: a shortened, more simple derivation of the items used by Zuwerink Jacks and Cameron (2003):
Source derogation - I think or say negative things about the person who wrote this text.
Asserting Confidence- I refuse to change my mind because I am confident in my opinion.
Emotional Affect - I get emotionally upset.
Selective Exposure - I tune-out the parts of the message I do not like.
Counterarguing - I think or talk about why the arguments in the text are faulty.
Social Validation - I think or talk about others who share my opinion about dementia.
Attitude Bolstering - I think or talk about the reasons behind my own beliefs about dementia.
Perceived knowledge (PK) was measured with the three perceived knowledge-items Zuwerink
Jacks and Cameron (2003) used. Items assessing perceived knowledge were (a) I consider myself to be
an expert on the issue of dementia, (b) I do not consider myself to be very knowledgeable about the issue of dementia (reversed), and (c) I have a lot of information about dementia. Responses to these
items were averaged to form a knowledge index (Cronbach’s α = .84), with higher numbers indicating greater perceived knowledge concerning dementia.
Prior knowledge (EN) was measured using five true/false-questions, to get an overall sense of
someone’s knowledge on this subject. Questions answered wrongly were coded 0 and questions answered correctly were coded 1. The responses to the questions were averaged to form a prior knowledge variable, ranging from 0 to 1. A lower score indicates a higher knowledge of dementia, a higher score indicates more knowledge.4
Attitude (AT) towards dementia was based on ten true/false-questions about dementia. The answers were coded: 1 for a plasticity answer, and 0 for a decline-answer. For each participant, the scores for the ten questions were averaged, to create an attitude variable, ranging from 0 to 1. A lower number on the AT-index indicates dominance of a decline-attitude towards dementia, while a higher number indicates dominance of a plasticity-narrative.5
The questions used to measure prior knowledge and attitude were based on online dementia ‘self-tests’ and quizzes, and common misconceptions about dementia, listed on the websites of Dementia Support Canada, the Alzheimer’s Association, Alzheimer’s Research UK, and the Dementia Forum. These questions were pretested in a small research colloquium meeting with researchers and research master students (n=10). Based on feedback from this group, one attitude-question was altered, since it seemed to elicit only decline-responses, regardless of existing attitude.
Demographic variables – Participants were asked to list their gender, educational level and
profession in commonly accepted ways. They were also asked whether there was someone close to them (friend or family member) suffering from dementia.
4.6.
Procedure
Data were collected online through Qualtrics,6 participants were directed to the survey via Prolific Academic. Firstly, participants read a short introduction and were asked for their informed consent. Next, they were presented with the three perceived-knowledge questions. On the next page, they filled in the five prior knowledge questions in random order, followed by a page where they were asked to answer the ten attitude questions, also in random order.
Then, Qualtrics randomly assigned participants to one of the four lists. Each participant was presented with a short vignette four times (see Materials). After each vignette, participants were presented with the seven possible reactions to the text and were asked to rate the likelihood of giving each of the responses on a seven-point scale (see Measures). Thereafter, several demographic questions were asked and there was room for participants to list any remaining remarks they had. Finally, participants were thanked for their participation and redirected to a Prolific Academic landing-page to collect their reward.
5 The questions used to measure ‘attitude’ can be found in the Appendix, section 9.2. 6 Qualtrics.com
17
4.7.
Analysis
To answer the research question and test the hypotheses, statistical analyses were conducted using IBM SPSS Statistics. Repeated measures 2x2 ANOVA’s and ANCOVA’s were conducted to assess the effects of the independent variables metaphor (metaphor/non-metaphor) and narrative (frame narrative: plasticity /decline) and potential moderators (attitude, perceived knowledge, prior knowledge) on the evoked total resistance and on the different resistance strategies. Differences were deemed significant at a .05 p-level.
Various respondents saw different versions of texts (4 lists) in different conditions. To account for the variation within stimuli, group mean centring7 was applied to the data, echoing Burgers et al. (2015), who conducted an experiment with a rather similar design.
7 Group mean centring concerns the process of transforming a variable by taking each score and subtracting it
from the mean scores (for that variable) for the group to which that score belongs (in this case the text list) (Field, 2009, p. 398-400; 829-830). Multilevel models with centred predictors tend to be more stable, and estimates from these models can be treated as more or less independent of each other (Field, 2009, p. 398-400; 829-830).
5. Results
Mixed factorial ANOVA’s and ANCOVA’s were conducted to assess the effects of the independent variables and potential moderators on the evoked total resistance and on the different resistance strategies. All direct effects were analysed with 2 (narrative: plasticity/decline) X 2 (frame: metaphorical/non-metaphorical) ANOVA’s generalizing across participants.
As stated in the previous chapter (4.7. Analysis), analyses are computed on the group-mean-centred dataset. However, for reader convenience, Table 1, 2, 3, 6, and 7 give the unadjusted means.
5.1.
Narrative
To assess the effect of the narrative presented in the experimental vignette, we look at scores for the seven resistance strategies and for the total resistance. The means and standard deviations can be found in table 1. Paired samples t-tests for each resistance type and for total resistance, only revealed a significant difference between text with a plasticity narrative and texts with a decline narrative for negative affect. The resistance strategy of negative affect was rated significantly more likely as a response to decline-texts, than to plasticity-texts (t (139)=-4.62, p=<.001). For all other strategies and for total resistance, no significant effects of narrative were found.
Table 1. Effects of text narrative (decline/plasticity) on the likelihood of employing different strategies to resist and total resistance (1=not at all likely, 7=very likely).
Narrative
Plasticity Decline Total
Source Derogation 2,00 (1,22) 1,94 (1,07) 1,97 (1,07) Assertion of Confidence 2,76 (1,55) 2,70 (1,51) 2,73 (1,47) Negative Affect 3,70 (1,61)* 4,13 (1,61)* 3,91 (1,51) Selective Exposure 2,46 (1,53) 2,44 (1,39) 2,45 (1,36) Counterarguing 2,81 (1,26) 2,80 (1,30) 2,81 (1,13) Social Validation 3,65 (1,45) 3,62 (1,44) 3,64 (1,35) Attitude Bolstering 3,82 (1,39) 3,79 (1,48) 3,81 (1,35) Total Resistance 3,04 (1,01) 3,05 (0,92) 3,04 (0,93) Means and standard deviations (between brackets), all effects are calculated over n=140. Significant differences (p<.05) between scores for the plasticity and decline narratives are marked with an asterisk*.
5.2.
Frame
To assess the effect of the type of frame (metaphorical/non-metaphorical) employed the experimental vignette, we look at scores for the seven resistance strategies and for the total resistance. The means and standard deviations can be found in table 2. Paired samples t-tests for each resistance type and for total resistance, revealed no significant effects of text frame.
Table 2. Effects of type of frame (metaphorical/non-metaphorical) on the likelihood of employing different strategies to resist and total resistance (1=not at all likely, 7=very likely).
Frame
Metaphor Non-metaphor Total Source Derogation 2,01 (1,27) 1,93 (1,05) 1,97 (1,07) Assertion of Confidence 2,73 (1,54) 2,72 (1,51) 2,73 (1,47) Negative Affect 3,86 (1,57) 3,96 (1,62) 3,91 (1,51) Selective Exposure 2,49 (1,51) 2,41 (1,39) 2,45 (1,36) Counterarguing 2,78 (1,34) 2,81 (1,23) 2,81 (1,13) Social Validation 3,61 (1,48) 3,66 (1,40) 3,64 (1,35) Attitude Bolstering 3,75 (1,47) 3,86 (1,41) 3,81 (1,35) Total Resistance 3,03 (1,02) 3,07 (0,90) 3,04 (0,93) Means and standard deviations (between brackets), all effects are calculated over n=140. Significant differences (p<.05) between scores for the metaphorical and non-metaphorical frame are marked with an asterisk*.
5.3.
Narrative x Frame
A repeated measures ANOVA was conducted to compare the effects of frame (metaphorical frame/ non-metaphorical frame) x narrative (plasticity/decline) on the resistance evoked by texts about dementia. Based on Levene’s test8, it was established that homogeneity of variances was not violated. Table 3 shows the means and standard deviations for the separate resistance strategies and for the total resistance. Test statistics, significance levels and degrees of freedom are displayed in table 2.
For total resistance, no significant main-effects of narrative (F(1,3)=1.03, p=.385) and frame (F(1,3)= 0,07, p=.805) were found, nor was there an interaction-effect of narrative x frame (F(1,3)= 0.34, p=.601). For each of the resistance strategies, separate repeated measure tests were conducted. None of the strategies showed significant effects for the independent variables and their interactions.
8 Maulchy’s M is not reported, since the test was conducted on two variables with two conditions per variable.
Table 3. Effects of type of frame (metaphorical/non-metaphorical) * narrative (decline/plasticity) on the likelihood of employing different strategies to resist and total resistance (1=not at all likely, 7=very likely).
Plasticity Decline
Met. Non-met. Met. Non-met.
Source Derogation 2,16 (1,63) 1,83 (1,23) 1.87 (1.35) 2,02 (1,29) Assertion of Confidence 2,76 (1,70) 2,77 (1,69) 2,70 (1,67) 2,69 (1,57) Negative Affect 3,56 (1,79) 3,83 (1,79) 4,16 (1,72) 4,10 (1,81) Selective Exposure 2,56 (1,78) 2,36 (1,60) 2,42 (1,64) 2,46 (1,53) Counterarguing 2,85 (1,66) 2,77 (1,41) 2,71 (1,60) 2,89 (1,52) Social Validation 3,66 (1,57) 3,64 (1,59) 3,56 (1,67) 3,69 (1,56) Attitude Bolstering 3,74 (1,58) 3,90 (1,56) 3,76 (1,66) 3,81 (1,53) Total Resistance 3,04 (1,14) 3,04 (1,01) 3,03 (1,05) 3,09 (0,92) Means and standard deviations (between brackets), all effects are calculated over n=140.
Table 4. Test statistics, p-values and degrees of freedom for the effects of the independent variables and their interactions on the likelihood of employing different strategies to resist and total resistance.
F p DF
Source Derogation Narrative 4,72 .118 1, 3
Frame 3,44 .161 1, 3 Frame*Narrative 1,86 .266 1, 3 Assertion of Confidence Narrative 0,31 .619 1, 3 Frame 1,08 .375 1, 3 Frame*Narrative 0,84 .427 1, 3
Negative Affect Narrative 0,08 .790 1, 3
Frame 1,08 .095 1, 3
Frame*Narrative 0,84 .464 1, 3
Selective Exposure Narrative 0,00 .985 1, 3
Frame 0,66 .476 1, 3
Frame*Narrative 1,09 .373 1, 3
Counterarguing Narrative 1,26 .343 1, 3
Frame 0,91 .411 1, 3
Frame*Narrative 0,40 .574 1, 3
Social Validation Narrative 0,65 .480 1, 3
Frame 2,02 .250 1, 3
Frame*Narrative 3,42 .162 1, 3
Attitude Bolstering Narrative 0,07 .806 1, 3
Frame 0,36 .592 1, 3
Frame*Narrative 1,00 .290 1, 3
Total Resistance Narrative 1,03 .385 1, 3
Frame 0,07 .805 1, 3
Frame*Narrative 0,34 .601 1, 3
5.4.
Moderator variables – Attitude, Perceived knowledge, Prior knowledge
Repeated measures ANCOVA’s were conducted too account for the effects of the potential moderator variables Attitude, Perceived knowledge, and Prior knowledge on the total resistance. Table 3 gives an overview of these analyses and shows no moderation of attitude, perceived knowledge and prior knowledge on total resistance.
Separate repeated measures ANCOVA’s were conducted to account for the effect of each moderator variable on the seven different resistance strategies. The prior knowledge index was used to create a categorical variable with two levels. Index scores below 0,5 were classified as low-knowledge. Index scores between 0,5 and 1 were classified as high-low-knowledge. 80,7% of all participants fell into the high-knowledge category. The remaining 19,3% fell into the low-knowledge category. No significant interactions were found when controlling for prior knowledge. Only significant outcomes for perceived knowledge and attitude are reported here.
Controlling for attitude only showed a significant interaction effect for the resistance type selective exposure of narrative x attitude(F(1,2)=78.73, p=.012). To further examine the effect of attitude on the scores for selective exposure, the attitude index (0-1) was used to create a categorical variable with two levels. Index scores below 0,5 were considered decline-attitudes. Index scores between 0,5 and 1 were considered plasticity-attitudes. Plasticity attitudes were highly dominant (93,6 of participants) over decline attitudes (6,4% of participants).
A repeated measures ANOVA was conducted for the selective exposure scores, the new attitude variable as predictor. Participants with a decline-oriented attitude showed more selective exposure to texts with both narratives (decline: M=2.38, SD=.66; plasticity: M=3.17, SD=1.60), in comparison to participants with a plasticity-oriented attitude, who showed less selective exposure to both texts with a decline narrative (M=2.41, SD=1.42) and texts with a plasticity narrative (M=2.41,
SD=1.53). Means and standard deviations are displayed in table 6.
Table 5. Test statistics, p-values and degrees of freedom for the effects of the potential moderator variables in interaction with the independent variables on the total resistance.
F p DF
Attitude Narrative*Attitude 0,64 .509 1, 2
Frame*Attitude 0,03 .890 1, 2
Frame*Narrative*Attitude 0,07 .820 1, 2
Perceived Knowledge Narrative*PK 0,15 .740 1, 2
Frame*PK 5,90 .136 1, 2
Frame*Narrative*PK 5,22 .150 1, 2
Prior Knowledge Narrative*PriorKn 0,01 .923 1, 2
Frame*PriorKn 1,99 .294 1, 2
Frame*Narrative*PriorKn 14,55 .06 1, 2
Table 6. Effects of attitude on selective exposure for narrative*attitude Decline attitude Plasticity attitude
Decline-text 2,83 (0,66) 2,41 (1,42)
Plasticity-text 3,17 (1,60) 2,41 (1,53)
Means and standard deviations (between brackets), all effects are calculated over n=140.
Controlling for perceived knowledge showed a significant interaction effect on negative affect for frame x perceived knowledge (F(1,2)=24.15, p=.039), and for narrative x frame x perceived knowledge (F(1,2)= 98,40, p=.010). To further examine the effect of perceived knowledge on scores for negative affect, the perceived knowledge index (0-7) was used to create a categorical variable with two levels. Index scores below 3,5 were considered ‘lower perceived knowledge’ (25% of participants fell into this category). Index scores between 3,5 and 7 were considered ‘higher perceived knowledge’ (75% of participants fell into this category). A repeated measures ANOVA was conducted for the negative affect scores, with the new perceived knowledge variable as predictor. For both frames and in all four conditions, higher perceived knowledge caused higher total resistance scores for negative effect that lower perceived knowledge. Means and standard deviations are displayed in table 7.
Table 7. Effects of perceived knowledge (PK) on negative affect for frame*PK and for narrative*frame*PK.
Frame*PK Narrative*Frame*PK Low Perceived Knowledge High perceived knowledge
Metaphor 3,75 (1,54) 4,20 (1,63) Decline*Met. 4,04 (1,73) 4,53 (1,67) Plasticity*Met. 3,46 (1,75) 3,89 (1,88) Non-Metaphor 3,86 (1,58) 4,27 (1,71) Decline*Non-Met. 3,95 (1,76) 4,54 (1,90) Plasticity*Non-met. 3,77 (1,77) 4,00 (1,86)
6. Discussion and conclusion
The aim of the current research was to explore the question: how much and what kind of resistance is evoked by metaphorical frames dealing with decline and plasticity narratives in dementia discourse, as opposed to literal frames presenting decline and plasticity narratives in this field, taking into account the effects of attitude towards dementia and both perceived and prior knowledge about this topic? An experiment was conducted to examine this question. The results are discussed below with regards to the hypotheses formulated in chapter 3.
6.1.
Narrative and Frame
No significant impact was found of frame metaphoricity on the amount and type of research evoked by the texts in different conditions. H1a: Metaphorical frames do not differ from literal frames when it
comes to the amount and types of resistance they evoke is therefore not confirmed. These results can
be seen as confirmation for H1b: Metaphorical frames differ from literal frames when it comes to the
amount and types of resistance they evoke. However, even though we might not have found a
significant difference between the metaphorical and non-metaphorical frames, our results do not match the effects found by Zuwerink Jacks and Cameron (2003). In their research, attitude bolstering and counterarguing appeared to be the strategies most likely to be used to resist persuasion. The results found in the present research do not indicate such a prevalence whatsoever. Moreover, the total resistance results reported in this study seem rather low, regardless of frame or narrative. Average of all measurements is a resistance score of 3.04, which is just slightly below the midpoint of 3.50 for the 7-point scale that was used. These numbers indicate that the texts read by participants evoked little resistance, but they could also point to the notion that people do not want to show resistance to these texts.
The topics used by Zuwerink Jacks and Cameron (2003) were the resistance-prone subjects of abortion and death penalty. These topics may be highly controversial, but presumably not as emotionally proximate for a lot of people, compared to the topic of dementia. In the case of dementia, there is not as much polarized, public debate as for abortion and death penalty. Dementia is not a topic which people would automatically link to discussion and resistance. Furthermore, in the case of capital punishment and abortion, there seems to be less direct and constant care and confrontation as opposed to dementia.
Dementia is an emotional and close topic for many people. Therefore their answers about the likelihood of resisting texts about this topic may originate from the desire to provide socially acceptable answers and not say anything negative about the illness and its patients.
One of the few significant effects found in our study was the effect of narrative on the use of the resistance strategy of negative affect. People proved to be more likely to use this strategy towards texts presenting a decline narrative as opposed to texts presenting a plasticity narrative. When reading the more negative decline texts, participants showed higher likelihood ratings of becoming emotionally upset. This effect might originate from the hopeless character of these decline texts: the negative image upsets more easily and might be something people do not want to hear, or in this case, read about.
6.2.
Attitude
One of the most surprising results of this study is the absence of a significant effect of participants’ attitude towards dementia. Both H2a Participants retaining a decline-oriented attitude towards
dementia show more resistance towards texts in which a plasticity narrative is propagated, regardless of metaphoricity of the frame, and H2b Participants retaining a plasticity-oriented attitude towards dementia show more resistance towards texts in which a decline narrative is propagated, regardless of metaphoricity of the frame cannot be confirmed. Overall, participants showed mainly
plasticity-oriented attitudes, but these attitudes did not affect the resistance shown towards texts opposing these attitudes.
The prevalence of plasticity-oriented attitudes is remarkable. Of the two distinct narratives distinguished by Frezza (In press), the decline narrative occurs much more frequently than the plasticity narrative. This discrepancy indicates a potential gap between the distribution of narratives within the discourse, and the attitudes existing in people’s mind. The high amount of plasticity-oriented attitudes could also derive from the urge to provide socially acceptable answers, not only to the resistance-questions but also to the attitude measurements. Additionally, instead of socially-accepted answers, the high plasticity-scores could also present a form of ‘wishful thinking’: besides the negative image present in discourse, people like to cling to the more positive narrative of plasticity.
Although selective exposure was one of the strategies rated as least likely to be used in the research of Zuwerink Jacks and Cameron, the results of our experiment indicated an effect of attitude on the resistance strategy of selective exposure. Even though the decline-oriented group of participants was rather small, this group did show significantly higher ratings of overall selective exposure towards the texts, meaning that they are more likely to tune out parts of messages they do not like. Their decline-oriented (more negative) attitude might cause them to be less hopeful overall and more motivated to evade messages that induce these hopeless feelings. These findings indicate that people displaying a decline-oriented attitude towards dementia might be harder to reach with messages about dementia since they tend to tune out.
These results demonstrate the necessity of searching for different ways to measure people’s attitude, to account for the display of socially accepted answers. It is also important to investigate whether and why there is a discourse-attitude gap, especially in light of Burgers (2016) findings that shifts of metaphor can be used to study societal shifts.
6.3.
Perceived and prior knowledge
The final hypothesis of this study was H3: The resistance evoked by metaphorically and
non-metaphorically framed texts advocating a decline or a plasticity narrative for dementia is moderated by (a) perceived knowledge, and (b) prior knowledge. This hypothesis cannot entirely be confirmed,
but some effects were found for perceived knowledge and prior knowledge.
When controlling for perceived knowledge, an effect on negative affect was found. Participants with high perceived knowledge about dementia – i.e. participants who think they know a lot about dementia – tended to show higher likelihood ratings of getting emotionally upset. The questions used to assess the perceived knowledge asked people whether they would rate themselves as experts and whether these people had a lot of information about dementia. We do not know whether these people viewed themselves as knowledgeable on an experience-level, or on a level of medical knowledge. The people with high perceived knowledge ratings might have also been the people for whom dementia is a topic closer to their hearts, which causes them to have a higher likelihood of displaying negative affect. Again, just like for selective exposure, negative affect is a strategy marked by Zuwerink Jacks and Cameron (2003) as ‘less likely’ to be used. They did not find significant effects of perceived knowledge on negative affect.
The results of this study showed no effect or prior knowledge. It could be argued that these questions were too general to display accurate effects. The majority of participants (over 80 per cent) fell under the ‘higher prior knowledge’ category. Besides, as stated above, knowledge about the topic of dementia needs to be specified further.
6.4.
Directions and limitations
The hypotheses that were set up to attempt to shed more light on the connections between metaphorical frames and resistance were hardly confirmed, nor rejected based on our results. This proofs that much further research is needed to look into the different aspects and influential factors in this field. Some limitations of the present research will be listed here, so future research can provide more accurate results. Furthermore, directions for further research will be provided.
A control-condition of the experimental material needs to be employed. Besides comparing literal and metaphorical frames, a non-framed version should be added as a baseline measurement.
With this additional baseline condition added, more insights can be gathered to contribute to the debate about opposing metaphor and literal language in framing research. Furthermore, more elaborate pre-testing of both the perceived- and prior knowledge and the attitude measurements needs to be conducted, to have a more solid image of what exactly is measured for the specific research topic. Lastly, it is important to note that the vignettes used in the present research are not authentic discourse: they were created for this study. Future research should try to account for this by either engaging in more elaborate pretesting of the experimental material or making use of actual discourse.
Besides these limitations, there are some main points that need to be considered in further research. Firstly, the differences between the outcomes of present research and the research conducted by Zuwerink Jacks and Cameron (2003), acclaim what has also been stated by Reijnierse et al. (2015): text themes seem to be of influence. It would be most helpful for the progress of research in the fields of resistance and metaphor research, to look at multiple themes, taking into account the characteristics of those themes. One very important theme characteristic to be considered is the (emotional) proximity of a topic to people. Differences in proximity may cause differences in the resistance strategies that are employed towards messages regarding particular topics. Different resistance strategies need to be countered or evoked in different ways. Lastly, socially accepted answers and strategies may differ across different themes and topics. It is important to create measurements that account for the possibility of socially accepted responses.
With regards to dementia perception in society, it is important to look further into the gap between narratives in dementia discourse and people’s personal attitudes towards dementia. To once again refer to the metaphors used by Louise Banks, the fictional linguist from the movie Arrival: further research is needed to find out how we can use language as a glue to bring peoples’ narratives closer together, and prevent language from turning into a weapon that stands in the way of a dementia-friendly society.
