1.1 What Is Synaesthesia and How Does It Relate to Creativity and Metaphor?

The above words were used by the Russian artist Wassily Kandinsky in 1895 to describe the powerful visual response that he had to a performance of Wagner’s Lohengrin at the Bolshoi Theatre in Moscow. For him, this was a life-changing experience that motivated him to abandon his career as a lawyer and embark on a career as an artist.

The ability to ‘see’ music that Kandinsky is describing here is a manifestation of a perceptual phenomenon, for which he is well known: synaesthesia. Synaesthetes (the term used to talk about people who have synaesthesia) have a particularly strong tendency to form cross-sensory associations (Reference SimnerSimner, 2007). For synaesthetes, who constitute approximately 4 per cent of the population (Reference Simner, Mulvenna and SagivSimner et al., 2006 – although estimates vary considerably across studies), the stimulation of one sensory pathway provokes involuntary stimulation of a different sensory pathway, leading people to make frequent associations between different senses (Reference HubbardHubbard, 2007). For example, certain sounds might be consistently associated with particular colours, textures might be associated with particular smells, tastes might be associated with particular colours, shapes or sounds, or days of the week as particular colours (Reference CytowicCytowic, 1989, Reference Cytowic1994). Synaesthetes may also form connections between different facets of the same sense; for example, reading black printed graphemes may trigger the perception of different colours (Reference Asher and CarmichaelAsher & Carmichael, 2013).Footnote ¹ Cutting across all these types of synaesthesia is a distinction between ‘associator synaesthesia’, where the associations occur ‘in the head’, and the less common ‘projector synaesthesia’, where the associations are projected or superimposed onto the stimulus that triggered the synaesthetic experience. So, for example, a grapheme-colour projector synaesthete will actually see the words appearing in different colours on the page, whereas an associator synaesthete will simply see the associations in their mind’s eye (Reference van Leeuwen, Simner and Hubbardvan Leeuwen, 2013).

The above quotation by Kandinsky is interesting. It is highly creative (music is not conventionally expressed as ‘wild crazy lines’), and it draws on metaphorical associations: the idea that high treble notes are associated with light, and low bass notes with darkness. Metaphor and creativity have been of central interest to the cognitive linguistic research community in recent years, but little is known about what propels people to use them. The aim of the study presented in this book is to identify and explore some of the clues that synaesthesia may provide to help us better understand the factors that drive people to make use of metaphor and creativity. In doing so, we take up the challenge proposed by Reference Ramachandran, Brang, Simner and HubbardRamachandran and Brang (2013, p. 1017), who suggest that:

Reference Cytowic, Simner and HubbardCytowic (2013, p. 403) suggests that ‘understanding synesthetic perception might help get a handle on the neurological basis of metaphor and creativity’. Our aim is to explore how the relationships between synaesthesia, metaphor and creativity play out in written language. The study is described in detail in Sections 2–5. In this section, we discuss the relationship between synaesthesia, metaphor and creativity from a theoretical standpoint, and introduce a number of other features of synaesthesia that may also form part of the puzzle: personification, empathy, evaluation and emotion. Our reasons for including these phenomena will become clear during the course of the section.

The heightened tendency to form cross-sensory mappings in synaesthetes also extends to the descriptions that they provide of sensory experiences. We can see this in the following three sentences, all of which were produced by synaesthetes after having been asked to write about sensory experiences that they enjoyed and sensory experiences that they disliked as part of the study we report on in this book:

In the first example, an auditory experience (the experience of listening to Bach’s music) is conveyed through a visual experience (the sight of a large, beautiful cathedral). In the second and third examples, an auditory experience is expressed by means of a tactile experience. These descriptions appear highly poetic, and are certainly not random or arbitrary. They could equally well have been produced by non-synaesthetes, but one of the questions that we explore in this book is whether synaesthetes have a heightened tendency to produce descriptions such as these, due to the neural characteristics of synaesthesia. We also explore the features of synaesthetes’ language that accompany these kinds of descriptions in order to shed light on the conditions that lead to their use.

Although some of the associations that synaesthetes form may seem random and arbitrary, there has been recognition in recent years that there is a degree of systematicity to the types of associations that synaesthetes make, leading some to suggest that there are shared ‘rules’ underlying synaesthesia (Reference Simner, Hubbard, Simner and HubbardSimner & Hubbard, 2013b). For example, when making sound-colour associations, synaesthetes tend to metaphorically map higher-pitched sounds onto lighter colours – interestingly, so do non-synaesthetes, although with less consistency (Reference Ward, Huckstep and TsakanikosWard et al., 2006). They also display a fairly consistent tendency to map higher-frequency words onto high-frequency chromatic colours (Reference SimnerSimner, 2007).

Indeed, the ability to form creative cross-sensory associations is not restricted to the synaesthete population. Descriptions of one sensory experience in terms of another (as in ‘a loud jacket’, ‘a sharp voice’ or ‘a soft sound’) have been identified in multiple languages around the world (Reference Speed, O’Meara, Roque and MajidSpeed et al., 2019; Reference WinterWinter, 2019). For Reference Marks, Simner and HubbardMarks (2013) these ‘synaesthetic metaphors’ are a hallmark of ‘weak synaesthesia’, a virtually universal ability to recognise similarities between different sensory domains which gives rise to such linguistic examples. For (‘strong’) synaesthetes, however, these associations may be considered to an extent ‘literal’ in that they are experienced perceptually and go beyond language – although, as we will see later in this section, the boundary may not be as clear-cut as such a definition may lead us to assume, and even in non-synaesthetes the cross-sensory associations that are formed have been shown to have a neurophysiological basis (Reference Ronga, Gola and ErvasRonga, 2016). However, it could be argued that synaesthetes’ connections ‘reflect to some extent the intuitive cross-modal correspondences of non-synaesthetes … synaesthesia relies not on specialised mechanisms specific only to synaesthetes, but to more universal mechanisms of general perception and cognition’ (Reference Simner, Simner and HubbardSimner, 2013, p. 160). Indeed, it has been proposed that all infants experience a form of synaesthesia due to the creation of ‘exuberant connections’ between neurons in the early stages of neural development, including those between areas of the brain that deal with distinct sensory systems. As the brain matures, some connections are strengthened and others ‘pruned away’; some cross-sensory connections remain, but most are lost or inhibited (Reference Maurer, Gibson, Spector, Simner and HubbardMaurer et al., 2013; Reference Maurer, Mondloch, Robertson and SagivMaurer & Mondloch, 2005). For synaesthetes, however, this ‘pruning’ or inhibition may not have taken place to such an extent as in non-synaesthetes. Brain imaging research has revealed that synaesthetes may exhibit stronger levels of connectivity between the areas of the brain involved in processing the senses they associate (see Reference Rouw, Simner and HubbardRouw, 2013 for a review), or that connections between these areas are somehow disinhibited (see Reference Mitchell, Simner and HubbardMitchell, 2013).

1.2 Metaphor and Its Relationship to Synaesthesia

The fact that synaesthetes appear to exhibit high levels of neural connectivity, and have a propensity to make associations between disparate entities, has led some synaesthesia researchers to draw links between synaesthesia and metaphor. As Reference Ramachandran, Hubbard, Robertson and SagivRamachandran and Hubbard (2005) suggest:

Reference Brang and RamachandranBrang and Ramachandran (2011) go on to propose:

Before exploring this link in more detail, it is worth taking a moment to consider what is meant by metaphor. Traditionally, metaphor has been described as a mechanism by which one entity is described in terms of another unrelated entity in a meaningful way, so for example, one might talk about ‘understanding’ in terms of ‘seeing’ (e.g. ‘I see what you mean’) or one might talk about ‘morality’ in terms of ‘cleanliness’ (e.g. ‘It was a dirty business’). In recent years, however, there has been increasing awareness of the fact that metaphor does not just involve describing one entity in terms of another but that at times it can involve an entity being experienced in terms of another. In other words, metaphor is often ‘experiential’ (Reference GibbsGibbs, 2017). This applies to many of the metaphors that people employ on a day-to-day basis which are based on physical interactions with the environment. For example, strong emotions are often described in terms of warmth or heat, our experience of time is talked about in terms of space and relationships are often understood in terms of proximity (Reference Lakoff and JohnsonLakoff & Johnson, 1980). Metaphors such as these are thought to underpin much of our ability to reason about abstract concepts and, for this reason, have been described as ‘primary’ metaphors (Reference GradyGrady, 1997, Reference Grady, Gibbs and Steen2019). There is increasing evidence to suggest that metaphors involving bodily experiences such as these are, to some extent, experienced in a physical way in that they provoke sensorimotor responses in the brain (Reference Boulenger, Hauk and PulvermüllerBoulenger et al., 2009). This means that, for example, a ‘weighty tome’, i.e. a book that literally feels heavy, will be deemed to contain more important information than one that feels light (Reference Chandler, Reinhard and SchwarzChandler et al., 2012; Reference Schneider, Rutjens, Jostmann and LakensSchneider et al., 2011), and hearing texture-based metaphors such as ‘a rough day’ or ‘a slimy person’ activates the texture-selective somatosensory cortex (Reference Lacey, Stilla and SathianLacey et al., 2012). The engagement of sensorimotor regions in response to metaphor has been found to increase when the metaphors are novel to the listener (Reference Cacciari, Bolognini and SennaCacciari et al., 2011; Reference Cardillo, Watson, Schmidt, Kranjec and ChatterjeeCardillo et al., 2012; Reference Desai, Binder, Conant, Mano and SeidenbergDesai et al., 2011). These findings suggest that when we are exposed to novel metaphors that make reference to sensorimotor experiences (e.g. ‘the flowers purred in the sun’ (Reference Cardillo, Watson, Schmidt, Kranjec and ChatterjeeCardillo et al., 2012)), we are more likely to experience them perceptually than when we are exposed to conventional metaphors that make reference to these experiences.

These findings support the theory of perceptual simulation (e.g. Reference BarsalouBarsalou, 1999), i.e. that hearing or reading a word triggers ‘mental simulations’ of perceptual or motor information that is stored in relation to that word. For example, hearing the word ‘lemon’ would trigger associations with the concept of ‘lemon’, including its taste, smell and appearance (Reference Gray and SimnerGray & Simner, 2015). It has been suggested that synaesthetes may experience these links in a more ‘conscious’ way, owing to the activation of neural connections which have been inhibited or ‘pruned’ in non-synaesthetes (Reference Gray and SimnerGray & Simner, 2015), and that these connections will then spread to other non-food-related words following phonological priming pathways until these too have tastes associated with them (Reference Simner and HaywoodSimner & Haywood, 2009). Thus, synaesthesia may be considered a ‘perceptual manifestation of implicit associations that lie at the heart of embodied cognition’ (Reference Gray and SimnerGray & Simner, 2015, p. 3).

Such a suggestion has important implications for the relationship between metaphor and synaesthesia, as it suggests a potential motivation for the connections synaesthetes make as mentioned above. Moreover, there is some evidence to suggest that core characteristics of ‘primary’ metaphorical mappings can also be found in the cross-sensory mappings that synaesthetes form. As we saw above, these ‘primary’ metaphors constitute the basic connections that exist between abstract experiences, such as ‘good’ and ‘bad’ and concrete environmental experiences such as ‘light’ and ‘dark’ (Reference GradyGrady, 1997). For example, when asked to associate shapes with sounds (Reference Spector and MaurerSpector & Maurer, 2008) and pitch with different levels of lightness (Reference Ward, Huckstep and TsakanikosWard et al., 2006) synaesthetes and non-synaesthetes tend to produce similar metaphorical pairings. Moreover, in primary metaphors, features such as intensity, valence and extremeness are mapped from source to target in systematic ways (Reference Barnden, Glasbey, Lee and WallingtonBarnden et al., 2003), and these principles can also be found in the cross-sensory associations that synaesthetes form involving sequences and space (Reference Jonas, Jarick, Simner and HubbardJonas & Jarick, 2013). Studies have shown that metaphorical relationships between time, number and space are motivated in the same way in both synaesthetes and non-synaesthetes but that the relationships are stronger in synaesthetes (Reference Cohen Kadosh and HenikCohen Kadosh & Henik, 2007).

A final relationship between metaphor and synaesthesia can be found in their origins. The formation of cross-sensory metaphors is often a non-conscious, automatic (i.e. spontaneous) and ‘pre-wired’ process, which is observable even in prelinguistic infants (Reference SeitzSeitz, 2005). Seitz identifies four types of metaphorical mappings that display these characteristics, namely (a) perceptual–perceptual (e.g. ‘a plate of spaghetti is a bunch of worms’), (b) cross-modal (i.e. synaesthetic) (e.g. red is a ‘warm’ or ‘hot’ colour), (c) movement–movement (e.g. the similarity between a spinning top and a dancing ballerina) and (d) perceptual-affective mappings (e.g. a piece of music is ‘cheerful’), arguing that, at least, in the initial stages of processing these metaphors operate largely outside of conscious awareness. Both metaphor and synaesthesia are thus pre-linguistic, automatic and involuntary (see also Reference CytowicCytowic, 2002).

1.3 Synaesthesia and Creative Metaphor

Synaesthesia brings with it benefits to creativity, which may be due to higher levels of neural connectivity, which are a significant common denominator to synaesthesia and creativity (Reference Faust and KenettFaust & Kenett, 2014; Reference Kenett and FaustKenett & Faust, 2019; Reference Rouw and ScholteRouw & Scholte, 2007). In other words, the higher levels of neural connectivity in synaesthetes facilitate an ability to meaningful associations between disparate entities, i.e. ‘associative fluency’. This is a central component of many creative thinking tasks (Reference Mulvenna, Simner and HubbardMulvenna, 2013). In her survey of the literature exploring the relationship between synaesthesia and creativity, Reference Mulvenna, Simner and HubbardMulvenna (2013) found extensive evidence for a synaesthetic advantage on objective measures of creative production that involve associative fluency. These included self-report measures of creativity and ‘divergent thinking’ tasks such as the ‘Torrance’ task (Reference TorranceTorrance, 1966), where participants are invited to think of as many creative uses as possible for everyday objects. The creative advantage shown by synaesthetes also manifests in creative activities such as the production of art and music (Reference Dailey, Martindale and BorkumDailey et al., 1997). Synaesthetes are disproportionately highly represented in fields such as these (Reference Rothen and MeierRothen & Meier, 2010), perhaps because their abilities to form cross-sensory associations help them produce creative images, films, animations and architectural designs (Reference Van CampenVan Campen, 2010, Reference Van Campen, Simner and Hubbard2013).

Associative fluency is also a key psychological process underpinning metaphor production (Reference Littlemore and LowLittlemore & Low, 2006). At this point, it is useful to return to the three examples given at the beginning of the section:

We have already identified the cross-sensory connections at play in each of these examples. The fact that one experience is being described in terms of another experience with which it is not normally associated, and that this process results in a new understanding of the original experience, means that all of these examples can also be said to involve metaphor (Reference GibbsGibbs, 1994). In describing the music of Bach as a ‘beautiful cathedral’ the writer of the first sentence is creating a metaphorical mapping between the internal structure of both the music and the cathedral, which draws attention to the fact that both combine an overarching beauty with an attention to detail. In describing a man’s voice in terms of warmth, the writer of the second sentence is drawing on the conventional metaphorical link between emotional attachment and heat. The writer of the third sentence is drawing on a conventional metaphorical mapping between pitch and sharpness.

These metaphors all involve a degree of novelty; we do not normally associate music with cathedrals, voices with the ability to melt our minds, or noises with red-hot needles. However, might they also be considered ‘creative’? In order to answer this question, we need to consider what ‘creativity’ means. For an idea to be considered ‘creative’, it must combine novelty with appropriateness (Reference CarterCarter, 2015). In other words, it needs to be both original and effective (Reference Runco and JaegerRunco & Jaeger, 2012). Creativity is therefore a two-part process; it involves a degree of free association, but there are also ‘rules’ governing the creative process which prevent random, meaningless combinations of ideas from being labelled as ‘creative’. The kinds of associations that synaesthetes make clearly meet the first of these two criteria in that they involve novel combinations between conceptual elements which have been previously unassociated (see Reference CarterCarter, 2015). However, are they always appropriate? In the examples above, even though the descriptions are unusual, they are grounded in conventional metaphorical mappings, combining pattern-reinforcing with pattern-forming aspects of creativity (Reference CarterCarter, 1999). For example, while the writer of the second sentence is drawing on the conventional metaphorical link between emotional attachment and heat, he or she also develops the metaphor in a creative way to say that his voice ‘melts his or her mind’. In previous work, we have shown that the creative use of metaphor often involves this kind of extension of an existing mapping (see Reference Fuoli, Littlemore and TurnerFuoli et al., 2021), and this kind of ‘language play’ has been shown to serve important communicative functions (Reference CookCook, 2000). The fact that the descriptions are grounded in existing mappings makes them accessible to non-synaesthetes. This characteristic seems to indicate a degree of audience awareness, as it enables the writers to explain their unusual experiences such that people who have not experienced them will understand. This renders them appropriate responses to the task. Indeed, the creative sentences produced by synaesthetes that we saw above are not wholly dissimilar to the kinds of creative cross-sensory imagery that we find in published poetry:

Although the core cross-sensory associations that synaesthetes form (e.g. that ‘Wednesday’ is ‘orange’) are not necessarily intentional or goal-driven (see Reference Ward, Huckstep and TsakanikosWard et al., 2006), examples such as those discussed above suggest that they give rise to elaborated, novel sensory experiences, the descriptions of which can be deemed ‘creative’. There is a degree of intentionality in the communication of these experiences. However, even if there had been no clear intention to be ‘creative’, the output could still be deemed so, as the role of intentionality in the creative process is a complex one. Many researchers highlight the role played by the unconscious mind in creative problem solving, and the fact that people ‘can often sense meaningful directions of exploration in trying to solve problems, even though they might not be aware of the actual reasons underlying these choices’ (Reference FinkeFinke, 1996, p. 388). The same arguments have been applied to metaphor. As Reference GibbsGibbs (2017, p. 90) argues, ‘many unconscious cognitive forces shape the online production and understanding of metaphors, which are simply not accessible to our conscious minds, despite our strong beliefs to the contrary’. It is therefore unlikely that there is a cut-off point between a cross-sensory metaphor that is produced as a result of a person’s deliberate decision to create one, and one that is simply the result of unconscious, automatic cognitive processes. Intentionality is not therefore a necessary prerequisite for creativity.

At this point, we can make the following interim claims:

• Synaesthetic associations and metaphorical mappings are alike in that they are, to a large extent, embodied, pre-linguistic and automatic;

• The enhanced connectivity in a synaesthete’s brain predisposes them to an ability to form connections between previously unrelated ideas. This is a hallmark of creativity, and also of metaphor;

• In writing about their experiences, synaesthetes seem to produce creative metaphorical mappings which draw upon embodied primary metaphorical connections.

A picture is therefore emerging of a relationship between metaphor, creativity and synaesthesia. Before introducing our study, however, there are two more areas which merit discussion due to their links with all three of these concepts: emotion and empathy.

1.4 Synaesthesia, Creative Metaphor and Emotion

Emotion has been shown to play a key role in inducing synaesthetic associations (Reference Callejas, Lupiáñez, Simner and HubbardCallejas & Lupiáñez, 2013), and synaesthetes often report strong emotional reactions to the cross-sensory associations that they form (Reference Cytowic and EaglemanCytowic & Eagleman, 2009). For example, they may derive pleasure from the associations that they form, but they experience strong negative emotional reactions to pairings that they perceive to be ‘incorrect’, such as when numbers are printed in the ‘wrong’ colour (Reference Callejas, Lupiáñez, Simner and HubbardCallejas & Lupiáñez, 2013; Reference Safran and SandaSafran & Sanda, 2015). This can be accounted for by the fact that synaesthetes have been found to exhibit hyper-connectivity between the amygdala (which is involved in the processing of emotions) and the sensory cortex (which is involved in the processing of sensory stimuli) (Reference Ramachandran, Hubbard, Robertson and SagivRamachandran & Hubbard, 2005). In other words, the fact that the sensory cortex is strongly connected to the amygdala suggests that emotion may play a role in the cross-sensory associations that synaesthetes form. In addition to this, people with auditory-visual synaesthesia show significantly more brain activation in the inferior parietal cortex during sound perception, than non-synaesthetes. Crucially, this part of the brain is responsible for multimodal integration, the perception of emotions and the interpretation of sensory information (Reference Neufeld, Sinke and DilloNeufeld et al., 2012).

Emotions can also play a part in the synaesthetic associations themselves. For example, some synaesthetes report unpleasant events in their lives being associated with a particular colour (Reference Safran and SandaSafran & Sanda, 2015), particular emotions may also trigger colour associations and words with particularly emotional connotations may be more likely to elicit synaesthetic responses (Reference WardWard, 2004). Emotions can also be the synaesthetic response to particular stimuli, as in the case of ‘tactile-emotion’ synaesthesia in which particular textures elicit emotional responses (Reference Ramachandran and BrangRamachandran & Brang, 2008).

Emotion is also heavily implicated in the production of cross-sensory metaphor in the non-synaesthete population (Reference Marks, Simner and HubbardMarks, 2013). Marks draws on ideas from embodied cognition to explain this link, arguing that metaphorical cross-sensory mappings in the non-synaesthete population emerge from ‘a nexus of affective and motor as well as perceptual responses to sensory stimulation’ (Reference Marks, Simner and HubbardMarks, 2013, p. 779). These include ‘physiognomic perception’, which he describes as ‘the relatively undifferentiated affective and expressive qualities of perceptual experience, as when we see a willow tree as “sad”’ (Reference Marks, Simner and HubbardMarks, 2013, p. 779). In other words, when we witness an object, we sometimes cannot help but experience an emotional response to it, and this response sometimes results in empathy. Indeed, evidence for a strong link between emotion and metaphor more generally can be found in neuroimaging studies (Reference Citron and GoldbergCitron & Goldberg, 2014; Reference Citron, Zervos, Baicchi, Digonnet and SandfordCitron & Zervos, 2018) which demonstrate that even conventional metaphorical sentences are more emotionally evocative than their literal counterparts. Affect and involvement have been shown to be a driving force in the production of metaphor (Reference Foolen, Foolen, Lüdtke, Racine and ZlatevFoolen, 2012), and personal emotional experiences have been found to lead people to produce creative metaphor (Reference Fainsilber and OrtonyFainsilber & Ortony, 1987). People have been found to use more metaphor, particularly creative metaphor, when talking about personal emotional experiences (Reference Williams-Whitney, Mio and WhitneyWilliams-Whitney et al., 1992), and it appears to be the first-hand experience of these emotions which provides the motivation for creative metaphor production (Reference MacCormacMacCormac, 1986). Other studies have shown that, even in non-synaesthetes, one of the main things that drive people to make metaphorical connections between colours, patterns and music is emotion (Reference Lindborg and FribergLindborg & Friberg, 2015; Reference Palmer, Schloss, Xu and Prado-LeónPalmer et al., 2013).

Furthermore, the instruction to think explicitly about how one might react in an emotional way to an imagined experience has been shown to trigger associative fluency, which in turn leads to the production of creative metaphor (Reference Lubart and GetzLubart & Getz, 1997). Lubart and Getz argue that experiencing an emotional reaction to a given concept or experience can help trigger a resonating endocept, whose attachment to the original concept or experience then becomes active in working memory. They argue that:

They illustrated this theory by having a group of business students focus consciously on the emotional experiences that they associated with being in an elevator, and then asked them to generate metaphorical associations that an elevator might have. This led to a number of associations, including a metaphor that was produced by one student, where being in an elevator was seen as being like an animal in a cage. The student then went on to reason that, like animals, people in an elevator may perceive their ‘cage’ to be boring because the scenery is unchanging. One idea resulting from this metaphor was to change the displays (e.g., posters) on the walls of elevators. Another insight from the zoo-cage metaphor was that people may find elevators uninviting because they lack features of their natural habitat; elevators could thus be improved by furnishing them in the style of a person’s living room. Thus, in this case, the instruction to access one’s emotional responses to the feeling of being in an elevator triggered the generation of an apt, creative metaphor along with its entailments.

Emotion has been shown to be a key driver in the creative process more generally (Reference RussRuss, 2013), and the use of multisensory imagery has been shown to be a key strategy that non-synaesthetes employ when invited to express their emotions creatively through the media of poetry and art, with embarrassment being found to provoke the use of tactile imagery, and romantic or sensual emotions provoking the use of olfactory and taste sensory images (Reference ShawShaw, 2008). This links to more general research showing that when people have negative emotional experiences, this drives the production of creative outputs (e.g. Reference BastianBastian, 2017; Reference SchubertSchubert, 1996), and in particular creative metaphor (Reference Fainsilber and OrtonyFainsilber & Ortony, 1987; Reference Littlemore, Turner and TuckLittlemore et al., forthcoming b; Reference WinnerWinner, 2018). This may be a reflection of the so-called ‘negativity bias’ which has been observed in the general population (Reference Jing-SchmidtJing-Schmidt, 2007). People have been found to give greater weight to negative entities, which means that they are more likely to pay attention to negative entities and events, and to remember them (Reference Rozin and RoyzmanRozin & Royzman, 2001). Negative experiences are therefore more vivid, and according to the vividness hypothesis (Reference Fainsilber and OrtonyFainsilber & Ortony, 1987), more vivid experiences trigger more creative metaphor use. Thus, the desire to produce creative metaphor may emanate in part from the need to share negative emotional experiences.

1.5 Synaesthesia, Personification and Empathy

An important characteristic that has been found to be strongly associated with synaesthesia is the tendency to personify objects (Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013). This characteristic, which is found in approximately 33 per cent of the synaesthete population (Reference Amin, Olu‐Lafe and ClaessenAmin et al., 2011), is automatic and involuntary (Reference Simner and HolensteinSimner & Holenstein, 2007), and can be viewed as a kind of metaphor as it involves perceiving an object or entity in terms of something that it is not (i.e. a human). The personification of stimuli allows social and affective characteristics to be attributed to them (Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013). For example, a synaesthete might attribute personality traits to linguistic inducers such as graphemes, an example of ‘ordinal linguistic personification’ (OLP) (Reference Simner and HolensteinSimner & Holenstein, 2007, p. 694). Reference Amin, Olu‐Lafe and ClaessenAmin et al. (2011) report that when synaesthetes attribute human characteristics to inducers such as these, they often refer to ‘human-style’ relationships between them, so inanimate objects might be ‘friends’, ‘enemies’, members of the same social group and so on. Reference Smilek, Malcolmson and CarriereSmilek et al. (2007) classified the social and affective characteristics that synaesthetes attribute to stimuli into four types: physical (gender, physical appearance); personal (cognitive ability, personality, mental states, moods, inclinations); relations (emotive and behavioural responses to other stimuli) and social role (occupation, familial and non-familial relationships).

While there is some controversy over the extent to which this ‘grapheme-personality synaesthesia’ is indeed synaesthesia as it does not involve a cross-modal perceptual experience, many definitions of synaesthesia have now been broadened to include non-sensory elements (such as personality) (Reference Sagiv, Sobczak-Edmans and WilliamsSagiv et al., 2017). Indeed, the mechanisms underlying the personification have been shown to be reminiscent of the more ‘prototypical’ synaesthetic connections described earlier, as they have been attributed to the hyper-connectivity that is a characteristic of a synaesthete’s brain. It has been postulated that the personification of linguistic inducers is caused by the heightened levels of ‘cross-talk’ between the left inferior parietal lobe (which is responsible for sequencing) and ‘social brain’ regions that are responsible for personality trait attribution (such as the amygdala and the somatosensory cortex) (Reference Simner and HubbardSimner & Hubbard, 2006).

The tendency to personify objects can also be found in the general population, particularly in infancy (Reference Sagiv, Sobczak-Edmans and WilliamsSagiv et al., 2017; Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013). According to the neonatal synaesthesia hypothesis (Reference Maurer, Mondloch, Robertson and SagivMaurer & Mondloch, 2005), all newborn babies experience synaesthesia, and the tendency to personify inanimate objects has also been attested in early infancy (Reference PiagetPiaget, 2002). Piaget argues that this form of thinking, which usually recedes in adulthood (although see Reference Degani, Onysko, Callies and DeganiDegani & Onysko, 2021), provides a mechanism for constructing reality with the self as the model. The brain areas associated with self-referential processing have also been found to be implicated in animistic thinking, including personification (Reference Sagiv, Sobczak-Edmans and WilliamsSagiv et al., 2017; Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013), and Reference Amin, Olu‐Lafe and ClaessenAmin et al. (2011) attribute the synaesthete’s heightened use of personification to heightened activation of the precuneus, which is associated with self-referential processing. Therefore, it may be reasonable to suggest that types of synaesthesia which involve personification are a ‘residual expression of childhood animism’ (Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013, p. 235), just as the more sense-based forms of synaesthesia may be a residual expression of the abundant neural connections present in early infancy. Interestingly, as with other kinds of synaesthetic connections, the principles underpinning the ways in which inanimate objects are personified have been found to be similar in synaesthetes and non-synaesthetes (Reference Sobczak-Edmans, Sagiv, Simner and HubbardSobczak-Edmans & Sagiv, 2013). For example, when asked what kind of personalities they would attribute to letters of the alphabet, both synaesthetes and non-synaesthetes attributed positive personal characteristics to frequently used letters and negative personal characteristics to less frequently used letters.

Once an inanimate object has been personified, many synaesthetes then go on to empathise with the object in some way, for example by expressing sympathy for an object that is excluded from the group (Reference Amin, Olu‐Lafe and ClaessenAmin et al., 2011). This has led some researchers to posit a relationship between synaesthesia and empathy levels, with mixed results. In one study of empathy levels in personifying synaesthetes, some participants were indeed found to exhibit higher levels of empathy than average, whereas others scored considerably lower (Reference Amin, Olu‐Lafe and ClaessenAmin et al., 2011). This finding led the researchers to suggest two different developmental routes to grapheme personification, ‘one representing an exaggeration or extension of normal processes (from the domain of human interaction to other domains), while the other could represent impaired criteria for personification, or sensitivity to the “wrong” cues’ (Reference Amin, Olu‐Lafe and ClaessenAmin et al., 2011, p. 277). However, mirror-touch synaesthesia, in which observing touch evokes a conscious sensation on the observer’s own body (e.g. Reference Blakemore, Bristow, Bird, Frith and WardBlakemore et al., 2005) entails ‘self-other blurring’ (Reference Maister, Banissy and TsakirisMaister et al., 2013), and has been linked with heightened emotional empathy and ability to read facial expressions of emotion (Reference Banissy and WardBanissy & Ward, 2007; Reference Ward, Schnakenberg and BanissyWard et al., 2018).

The existence of mirror-touch synaesthesia provides further evidence for the link between synaesthesia and embodied cognition. In Section 3, we introduced the theory of embodied cognition, noting that some researchers have suggested that synaesthetes consciously experience embodied links between perceptions and words – that synaesthesia is therefore a case of ‘disinhibited embodiment’ (Reference Gray and SimnerGray & Simner, 2015, p. 1). In synaesthetes and non-synaesthetes alike, mirror neurons allow for understanding, imitation and prediction of the behaviour of those around us by simulating these actions. However, when watching an action leads to a physical sensation related to that action (as in the case of mirror-touch synaesthesia), it may be that this is due to the synaesthete consciously experiencing this embodied simulation that is found in the population as a whole (Reference Banissy and WardBanissy & Ward, 2007; Reference Gray and SimnerGray & Simner, 2015). Reference Banissy and WardBanissy and Ward’s (2007) finding that mirror-touch synaesthetes exhibited higher levels of emotional empathy could also indicate that mirror neurons have a role to play in empathy.

At this point, we can suggest a potential role for empathy in the overall picture we have built up over this section. The studies introduced here allow us to propose potential links between synaesthesia, creativity, metaphor and empathy. It is these links that we aim to explore in more depth over the course of this book.

1.6 Bringing It All Together: Our Study

We have seen that synaesthesia, metaphor and creativity are interlinked, and that emotion and evaluation can modulate these relationships. We have also seen that personification, as well as being a type of metaphor in itself, is an important feature of many synaesthetes’ connections. Specifically, we have seen that: (a) synaesthetes display higher levels of creativity than the general population; (b) the cross-sensory associations that synaesthetes make appear to be metaphorically motivated, and may also be considered creative; (c) the neurological architecture of the synaesthete is characterised by higher levels of connectivity between parts of the brain that are responsible for sensory awareness, emotion and self-referential processing; (d) synaesthetes’ cross-sensory associations are often associated with particularly high levels of emotion, personification and empathy, and (e) emotion and evaluation are closely linked and the need to express emotional experiences or evaluation has been found to drive the production of (creative) metaphor. These attested links between synaesthesia, associative fluency and emotion provide support for the idea that it is the strong, emotional responses to sensory experiences that synaesthetes have that drive their creative use of metaphor.

It may be useful to explore the relationship between these characteristics in order to better understand the role that they play in metaphorical creativity and potentially in creativity more generally. As we saw at the beginning of this section, Reference Ramachandran, Brang, Simner and HubbardRamachandran and Brang (2013, p. 1017) suggest that:

We take up their challenge by conducting an in-depth examination of what ‘looks like’ creative metaphor, and identifying features that appear to be associated with this style of thinking. By doing so, we hope to shed light on the factors that seem to trigger the production of creative metaphor in synaesthetes and, by extension, in the population more generally. A key part of this enquiry involves identifying key characteristics of synaesthetes’ descriptions of positive and negative sensory experiences, and exploring the ways in which they differ from descriptions provided by non-synaesthetes. We explore similarities and differences between people with and without synaesthesia in terms of (a) their propensity to produce cross-sensory creative metaphor when describing positive and negative sensory experiences, (b) the clusters of features that accompany or trigger the production of cross-sensory creative metaphor, and (c) the associations they provide in response to emotion words. Our overall aim is to establish the extent to which the responses provided by synaesthetes resemble those made by non-synaesthetes in both qualitative and quantitative terms.

We explore the metaphorical nature of the associations that synaesthetes make, compare their propensity to make such associations with that of non-synaesthetes and identify the factors that co-occur with, or seem to give rise to, these kinds of associations, with a particular focus on evaluation, empathy and personification. We seek to establish the extent to which the associations that are produced by synaesthetes involve metaphor, whether or not synaesthetes form creative metaphorical mappings more often than the general population, and if so, what (if anything) synaesthesia might tell us anything about associative fluency and metaphorical creativity. To put it another way, our aim is to conduct an in-depth examination of a style of thinking that ‘looks like’ creative metaphor, to identify features that appear to be associated with this style of thinking in order to better understand the nature of metaphorical creativity.

Our research questions are therefore as follows:

1. 1. How do the responses of synaesthetes and non-synaesthetes differ from one another:

   1. a. When asked to write critical descriptions of positive and negative sensory experiences?

   2. b. When asked what words they associate with emotion words?

2. 2. What do these differences reveal about the relationship between metaphor, creativity and emotion?

In order to answer these questions, we compare and contrast the responses provided by the synaesthetes and non-synaesthetes in terms of:

• The range and types of cross-sensory associations that are made

• The extent to which they produce (creative cross-sensory) metaphor

• References to personification, empathy, emotion and other features emerging from the data

• The extent to which and the ways in which all of the above interact with one another

Our hypotheses are that in comparison with non-synaesthetes, synaesthetes will make use of a wider range of cross-sensory associations, and make more use of cross-sensory associations, personification, empathy and emotion. We also hypothesise that that there will be more inter-relationships between these response types in the synaesthete population than in the non-synaesthete population due to the higher levels of neural connectivity discussed in this section. The study also has an exploratory element in that we do not restrict ourselves to these response types and aim to identify, through a maximally inclusive iterative coding procedure, other response types that have not been identified in the literature.

In the remaining sections, we outline the methodology used in the study and present the findings. In Sections 2 and 3, we describe the methodology (Section 2) and discuss the findings (Section 3) from the first part of our study, in which we investigated the responses provided by synaesthetes and non-synaesthetes when asked to describe something they liked and something they did not like to see, hear, taste, smell and touch. In Section 4, we present the methodology and discuss the findings from the second part of our study in which we explored the responses that our participants produced when presented with emotion words. In Section 5, we provide a brief conclusion and consider the implications of our findings for our understandings of synaesthesia, metaphor and creativity.

2.1 Introduction

The aim of the study presented in this book is to identify and explore some of the clues that synaesthesia may provide to help us better understand creativity, or, more specifically, creative metaphor. In order to do this, we compare the responses that synaesthetes and non-synaesthetes provide to sensory and emotive prompts.

Through the use of an online survey, developed using and hosted on the Qualtrics platform, we asked a group of people who identified as synaesthetes and a group of people who did not identify as synaesthetes to participate in two tasks in which they responded to prompts involving sensory experiences and emotion words. For the sake of brevity, the two groups are labelled ‘synaesthetes’ and ‘non-synaesthetes’, although we acknowledge that the distinction is by no means as clear-cut. The questionnaire is shown in Figure A1 in the additional resources.

2.2 Participants

All participants were recruited online. Synaesthete participants (N=20) were recruited through the UK Synaesthesia Association (who included a link to our study in their member’s newsletterFootnote ²), the Synesthesia ListFootnote ³ and the synaesthesia subreddit.Footnote ⁴ The URL was also shared on Twitter, using the hashtags #synesthesia, #synaesthesia and the URL link was shared to synaesthesia-related groups on Facebook. The non-synaesthetes (N=20) were recruited online through social media via an advertisement. The link was open for twelve weeks in total. Participants in the non-synaesthete group were advised that if they had synaesthesia they should not complete the questionnaire. This recruitment procedure meant that one of our groups of participants had self-reported a strong tendency to form cross-sensory associations, and the other group had not.

In the advertisement, potential participants were told that the research would involve a series of short writing tasks, but no mention was made of metaphor or creativity, as this may have led to response bias (see e.g. Reference FurnhamFurnham, 1986). The Participant Information Sheet indicated that the purpose of the study was to ‘investigate the link between language and synaesthesia’. All participants were native-level speakers of English and over eighteen years of age. The study was granted ethical approval by the University of Birmingham. The demographic breakdown of our participants (age and gender) is shown in Table A1 in the additional resources.

Synaesthete participants were asked what kind(s) of synaesthesia they had and whether they had associator or projector synaesthesia. Seventeen reported experiencing associator synaesthesia, one reported experiencing projector synaesthesia and two reported experiencing both types of synaesthesia. Participants reported a range of types of association, with most participants reporting more than one type. The types of synaesthetic association reported are shown in Table A2 in the additional resources.

This research design is not without its risks. Firstly, the participants in the ‘synaesthete’ group may have contained a number of ‘false positives’, i.e. individuals who would not be classified as having synaesthesia as measured by standard tests (see Reference Mulvenna, Simner and HubbardMulvenna, 2013). However, to the best of our knowledge standard tests for synaethaesia, such as the ‘Test of Genuineness’ (Reference Asher, Aitken, Farooqi, Kurmani and Baron-CohenAsher et al., 2006), focus exclusively on colour synaesthesia (Reference Carmichael, Down, Shillcock, Eagleman and SimnerCarmichael et al., 2015) and in our study we were interested in a wider range of synaesthetic experiences. The main focus of our study was on the tendency to make cross-sensory associations and characteristics that appear to accompany that tendency.

Secondly, the self-selecting nature of this group may have meant that they saw the task as an opportunity to ‘perform’ their identity as synaesthetes. This may have led to them providing more detailed responses or responses that demonstrate features that they thought we might be looking for in the project. In order to mitigate this, as we saw above we made no mention of metaphor or creativity in the recruitment or in the instructions for participants.

2.3 Procedure

In the first task, participants were asked to identify and describe something that they did and did not like to see, hear, taste, smell and feel. There was no word limit or time limit for this task. We (the two authors) coded their free-text responses for features that have been observed in the literature (e.g. personification, empathy, references to emotion, and metaphor) as well as other features that emerged from our analysis. We saw in Section 1.2 that synaesthetes have been found to perform better than non-synaesthetes in divergent thinking tests, which are thought to be measures of creativity. However, these tests are largely decontextualised and generally do not require participants to discuss real-world experiences or to provide reasons for their responses. To the best of our knowledge, no studies to date have looked at metaphorical creativity when synaesthetes and non-synaesthetes are performing free-writing tasks in which they are invited to provide personal responses to sensory experiences. Moreover, given the apparent importance of emotion, personification and empathy in synaesthetic responses to stimuli, we were interested in comparing their reactions to phenomena that they did and did not like. As we saw above, there is evidence to suggest that people are particularly likely to produce creative metaphor when reacting in an emotional way to a stimulus (Reference Lubart and GetzLubart & Getz, 1997), and this effect is intensified when they are responding to negative emotional experiences (Reference Fainsilber and OrtonyFainsilber & Ortony, 1987; Reference WinnerWinner, 2018). We were interested in investigating whether this valence extends to the synaesthete population, as they have been found to express strong positive and negative emotional reactions to sensory experiences.

The second task was a word association task in which we investigated their responses to the six ‘basic’ emotion words: ‘happiness’, ‘sadness’, ‘fear’, ‘anger’, ‘surprise’ and ‘disgust’ (Reference EkmanEkman, 1992). Whereas in the first task, we hoped to elicit personal, potentially emotional responses to external entities, in this task we sought to investigate how synaesthetes and non-synaesthetes respond to emotional language by identifying what kinds of things they associate with it. Again, we coded their responses for features that have been identified in the literature on synaesthesia (e.g. colours, letters of the alphabet, physical sensations, novelty) as well as other features that emerged in our analysis of the data (e.g. scenarios, actions, physical objects).

In both tasks, we identified the ways in which the synaesthetes and the non-synaesthetes differed from one another in terms of their response patterns, and examined the different ways in which different types of response interacted with one another in both populations in order to better characterise the response patterns exhibited by the synaesthetes. Our hypothesis was that these response patterns would provide insights into the mechanisms underlying synaesthetes’ purported higher propensity towards the production of creative metaphor.

In this section, our focus is on the first task. We describe the methodology, outline the characteristics of the corpus and introduce the coding scheme that was used to annotate it.

2.4 Characteristics of the Corpus

Each participant produced a total of 10 written responses (a positive and a negative response for each of the five senses), resulting in a total of 400 written responses (N=40 participants x 10 responses each). These written responses resulted in a corpus made up of 15,151 words for the synaesthete responses and 6,970 words for the non-synaesthete responses.Footnote ⁵ In Table A3 in the additional resources we provide a breakdown of the corpus, showing the number of words that were used to talk about each of the senses in positive or negative terms.

It is interesting to note that the synaesthetes wrote significantly more than the non-synaesthetes. This may be because the synaesthetes’ sensory experiences are to a certain extent ‘richer’ than those of their non-synaesthete counterparts as not only do they experience the stimulus through the primary sense, they also have access to the associations this stimulus triggers. As we saw in Section 1, synaesthesia may also entail a greater emotional engagement with the experience (e.g. Reference Cytowic and EaglemanCytowic & Eagleman, 2009). They also demonstrate higher performance in a variety of memory tasks (see Reference Meier, Rothen, Simner and HubbardMeier & Rothen, 2013 for a review). It has been suggested that this is due to the increased richness of their experience; because synaesthetes encode additional features into their memory (e.g. colour may be associated with numbers), they can take advantage of a wider range of retrieval cues (Reference Rothen, Berry, Seth, Oligschläger and WardRothen et al., 2020) which may allow them to recall richer, more emotional experiences in greater detail. Although the longer productions by the synaesthetes may partly be explained by the fact that the synaesthetes’ sensory experiences may have been richer than those of the non-synaesthetes, they may also have been explained by the fact that they may have had a correspondingly strong urge to communicate these rich experiences. The longer responses may also reflect the characteristics of the sample of synaesthetes used in our study. The fact that they were recruited through a synaesthesia forum suggests that they have strong identities as synaesthetes and that they may have viewed the activity as an opportunity to express these identities.

There was also a difference in behaviour relating to the two groups’ descriptions of positive and negative sensory experiences. Whilst both groups said more about positive sensory experiences than negative sensory experiences, the difference was noticeably more marked in the synaesthetes than in the non-synaesthetes. We also found that the synaesthetes produced more words in response to positive prompts for sight, sound, smell and touch and more words in response to negative prompts for taste. Non-synaesthetes produced more words in response to positive prompts for sight, sound and smell, and more words for negative prompts for taste and touch. These findings suggest that the ‘negativity bias’ discussed in Section 1 may be less marked in synaesthetes than in non-synaesthetes, possibly reflecting the levels of ‘enjoyment’ that synaesthetes sometimes report when making cross-sensory associations.

2.5 Coding Procedure and Scheme

Both authors coded the entire dataset together and engaged in extensive, in-depth discussion in order to establish the coding categories. The decision to work together on the entire dataset rather than taking a section each and comparing was due to the fact that we were not working exclusively with a priori categories. A major part of the research endeavour involved developing the scheme either through identifying new categories or by delineating categories that reflected or built on ideas that had been suggested in the literature on synaesthesia. This involved making multiple passes of the data and discussing the emerging structure of the scheme as the categories became apparent.

Responses were first coded according to the sense and valence of the sensory experience that was being described (e.g. responses to ‘write about something you like to hear’ were coded as ‘sound’ and ‘positive’). Following this, a coding scheme was developed to identify the types of responses that were provided. This scheme was first informed by previous research into creativity and synaesthesia as introduced in the literature review, and then honed through an iterative process. It had three overarching groups of categories: response sense, to identify the senses activated in the responses; response type, to identify the broad types of response being given, e.g. emotional, empathic, etc.; and ‘metaphorical’ theme. A single response could be coded into more than one category and more than one group.

3.1 Introduction

As we saw in Section 1, the overarching aim of the study described in this book is to explore the nature of ‘creativity’, specifically metaphorical creativity, by examining it through the lens of synaesthesia. We seek to identify factors of synaesthetes’ responses that could be considered ‘creative’ or may even motivate creativity, and to compare their responses to those produced by non-synaesthetes.

In Section 2, we described the methodology employed in the first part of our study. We found that synaesthetes produced much longer responses than non-synaesthetes and suggested that this difference may reflect the fact that synaesthetes’ sensory experiences are in some way ‘richer’ than those of their non-synaesthete counterparts as not only do they experience the stimulus through the primary sense, they also have access to the associations this stimulus triggers.

In this section, we go into more detail on the nature of responses given by synaesthetes and non-synaesthetes when describing positive and negative sensory experiences. We start by providing a flavour of the ways in which synaesthetes and non-synaesthetes differ in terms of the use they make of the senses in their responses. We then move on to our exploration of the differences between the two groups in terms of use that they made of the response types that were described in Section 2. We focus in particular on the metaphorical themes that they draw on.

Tables showing data on the responses produced for each sense, split by valency, can be found in the additional resources. Tables A4 and A5 show the sense responses produced by synaesthetes and non-synaesthetes respectively. Tables A6 and A7 show the response types produced by synaesthetes and non-synaesthetes respectively. Tables A8 and A9 show the metaphor themes produced by synaesthetes and non-synaesthetes respectively. Tables A10 and A11 show the correlations between the response types produced by synaesthetes and non-synaesthetes respectively.

3.2 How Do the Synaesthetes and the Non-synaesthetes Compare with One Another in Terms of the Senses That They Employ in Their Responses?

Given that a key characteristic of synaesthesia is a tendency to describe one sensory experience in terms of another, one would expect to find more references to the senses overall, and more cross-sensory mappings in the descriptions provided by the synaesthetes than in those provided by the non-synaesthetes. We therefore looked at the range of sense types involved in the descriptions provided by each group and investigated the extent to which they cross sensory boundaries when describing their experiences. We also investigated whether the senses used by the synaesthetes in their descriptions resembled those used by the non-synaesthetes, and the extent to which their response patterns resembled those that have been reported in the literature on cross-sensory associations.

Our results are shown in Tables A4 and A5 in the additional resources, and are summarised in Tables 1 and 2.

As expected, synaesthetes drew on the senses a lot more than the non-synaesthetes, (210 times versus 61), and they made use of a wider range of senses, often when describing a single sensory experience. In the following example, a synaesthete draws on taste, smell, sound, sight and touch to describe the experience of eating bananas and honey:

Meanwhile, the non-synaesthetes did produce cross-sensory mappings, but these were largely limited to the sense of touch, which coheres with corpus-based studies showing that touch words are inherently more ‘multisensory’ than words relating to the other senses (Reference Strik LieversStrik Lievers, 2015).

Some researchers investigating the use of cross-sensory associations in the general population have suggested that there is a ‘hierarchy’ of the senses, with some senses being more likely than others to be used as the source of cross-sensory associations. The earliest proponents of this idea were Reference UllmannUllmann (1967) and Reference WilliamsWilliams (1976), who proposed that ‘touch’ is the sense that is most likely to be used to talk about other sensory experiences, followed by taste and then smell, with sight and sound being most likely to be the target of the description. Although there is some empirical support for this hierarchy (Reference DayDay, 1996; Reference ShenShen, 1997; Reference Shen and CohenShen & Cohen, 1998; Reference Strik LieversStrik Lievers, 2015; Reference WiseWise, 2003), it has been critiqued in a recent corpus-based study by Reference WinterWinter (2019), who questions the idea of a monolithic hierarchy and argues for a more nuanced and guarded approach. Interestingly, however, the non-synaesthetes’ responses in our study bore a much closer resemblance to Reference WilliamsWilliams’ (1976) hierarchy than the synaesthetes’ responses.

To sum up, synaesthetes and non-synaesthetes differed considerably in their use of sensory responses. Stimulus-coherent responses were rare in both groups, indicating that cross-sensory mappings are not restricted to synaesthetes. However, synaesthetes drew on a much wider range of senses in their responses than non-synaesthetes. They were also less likely than the non-synaesthetes to adhere to any kind of hierarchy in their cross-sensory mappings.

We also observed differences in the behaviour of the synaesthetes and non-synaesthetes in terms of the valence of sensory experiences which attracted sense-based responses. These differences are shown in Table 3.

Synaesthetes were more likely to draw on the senses when talking about positive items than when talking about negative items, whereas this tendency was reversed for the non-synaesthetes. A chi-square test comparing both groups’ use of senses in their responses to both positive and negative stimuli showed this difference to be statistically significant χ² (1, N = 40) = 4.86, p = .03. These findings suggest that for the synaesthetes, sensory responses form part of their usual way of conceptualising and processing experiences, both positive or negative, whereas for non-synaesthetes sensory responses are more likely to be used for negative experiences. However, it should be noted that this propensity seems largely driven by the ‘taste’ sense, which was used for negative responses almost twice as much as for positive responses.

We have seen in this section that when describing sensory experiences, synaesthetes differed considerably from non-synaesthetes in terms of the frequency with which they made cross-sensory associations, the senses that they drew on and the range of senses employed. They had much more to say about the experiences, especially the positive ones; they were more likely to make cross-sensory associations; they employed a wider range of senses; and their responses did not follow the hierarchy of the senses model as closely as the non-synaesthetes. The synaesthetes’ propensity to make connections between a wider range of senses than non-synaesthetes could be said to constitute a form of associative fluency which, as we saw in Section 1, is involved in creative production. According to Reference Fainsilber and OrtonyFainsilber and Ortony’s (1987) ‘inexpressibility hypothesis’, the need to describe an experience that is not shared tends to attract more metaphor as a means to make one’s message understood, and it is reasonable to assume that these metaphors may also be more ‘creative’ due to a lack of conventional mappings to draw upon. On the other hand, the senses that the non-synaesthetes draw on are more likely to follow conventional patterns, and to cohere with research into the hierarchy of the senses. The fact that these are conventional means that they do not require further elaboration or explanation, and may therefore be less likely to provoke creative output. These sensory directions motivate a large number of conventional metaphors (e.g. smooth jazz). In the next section, we dig more deeply into the types of responses provided, with a view to exploring the extent to which the synaesthetes do indeed produce more ‘creative’ responses to express their experiences as expected.

3.3 How Do the Synaesthetes and the Non-synaesthetes Compare with One Another in Terms of Their Response Types?

In this section, we turn to one of the core questions underpinning our study: what do synaesthetes do when asked to describe positive and negative sensory experiences that non-synaesthetes do not do, and what might this tell us about the factors that trigger the production of creative metaphor? In order to answer this question, we look at the types of response that were triggered by the different sensory experience prompts, explore differences according to the valence of the stimuli and investigate the ways in which the synaesthetes differed from the non-synaesthetes in terms of their response types. Here we compare synaesthetes with non-synaesthetes in terms of their tendency to produce the different response types introduced in Section 2, and look at their answers in detail to help understand how the more general behaviour of the synaesthetes might help explain their greater use of creative metaphor. The numbers of each type of response produced by each of the two groups are shown in Tables A6 and A7 in the additional resources, and summarised in Table 4.

We can see from Table 4 that the synaesthetes produced far more tokens of these response types overall (556) than the non-synaesthetes (318). This is to be expected, given the fact that the synaesthetes produced longer texts overall than the non-synaesthetes. We can also see that synaesthetes were more likely than non-synaesthetes to produce responses involving metaphor, cognitive effects, emotional effects, empathy, personification, value judgements, physical effects and hyperbole. In contrast, non-synaesthetes were more likely than synaesthetes to produce responses involving literal explanations and associations. In order to establish whether any of these responses were significantly more likely to be favoured by one of the groups, we conducted Mann-Whitney U Tests.Footnote ⁶ Independent variables were the two groups, and dependent variables were the response types. The results are shown in Table 5.

The response types provided by the two groups did not differ significantly in terms of the valence of the sensory experiences. Both groups demonstrated similar patterns of response types in response to negative and positive sensory experiences, and where differences are observed (e.g. in the ‘personification’ response type, where synaesthetes seem to produce personification more in response to positive sensory experiences whereas non-synaesthetes produce it more in response to negative sensory experiences), the numbers are too small to be able to generalise. However, one point of interest here is that these findings differ from our findings regarding the senses that participants drew on in their responses, where we saw a marked difference in the use of senses across the two groups. As we saw above, synaesthetes produced noticeably more examples of sensory responses when describing positive sensory experiences whereas for the non-synaesthetes, the difference was less marked. For response type, however, both groups produced more responses for negative sensory experiences than for positive sensory experiences. It seems that, when examined through the lens of response type rather than response sense, the importance of valence is less marked. However, it is interesting to note how the two groups differ in their behaviour. While both groups produced more response types for negative sensory experiences, the trend towards negativity was slightly stronger in the non-synaesthetes than for the synaesthetes. Again, this suggests that the negativity bias discussed in Section 1 may be slightly weaker in synaesthetes than in non-synaesthetes.

We were also interested to see whether there were any significant relationships between the different response types within each of the two groups. The existence of such relationships might suggest that these processes are somehow linked. We therefore ran a series of Spearman correlation tests across all major response types for each group of participants separately.Footnote ⁷ We hypothesised that there would be more inter-relationships in the responses produced by the synaesthetes, given their heightened levels of neural connectivity. The findings from these analyses are shown in Tables A10 and A11 in the additional resources. As expected, there were far more significant correlations between the variables in the synaesthetes’ data than in the non-synaesthetes data. For the synaesthetes, there were significant positive correlations between: cognitive effects and metaphor (ρ = .24); cognitive effects and emotional effects (ρ = .29); emotional effects and physical effects (ρ = .28); emotional effects and empathy (ρ = .17); emotional effects and metaphor (ρ = .19); empathy and physical response (ρ = .26); ‘hyperbole’ and personification (ρ = .15); value judgements and metaphor (ρ = .17); metaphor and personification (ρ = .18). There was also a significant negative correlation between literal explanations and physical effects (ρ = −.16). For the non-synaesthetes, however, there were only two significant positive correlations. These were between cognitive effects and physical effects (ρ = .17), and cognitive effects and personification (ρ = .25). There were also three significant negative correlations, where literal explanations were shown to negatively correlate with cognitive effects (ρ = −.14), emotional effects (ρ = −.15), and physical effects (ρ = −.15). This finding may suggest that when describing emotional, physical or cognitive effects, literal language is inadequate, which provides further support to the growing body of evidence that metaphor is a crucial communicative tool when describing personal, embodied emotional experiences (see Reference Colston and GibbsColston & Gibbs, 2021; Reference Littlemore and TurnerLittlemore & Turner, 2019; Reference Semino, Swann, Pope and CarterSemino, 2011).

We can see from these findings that the responses of the synaesthetes were characterised by higher levels of metaphor and emotional, cognitive and physical effects, and, to a lesser extent, empathy, personification, value judgements and hyperbole than those of the non-synaesthetes, and the response types were more interrelated in the synaesthetes. It is particularly interesting to note that in the synaesthetes, three of the response types (cognitive effects, emotional effects and value judgements) correlate with metaphor, suggesting that the desire to communicate experiences such as these may trigger the production of metaphor. Thus, a picture is beginning to emerge of the ways in which synaesthetes respond when asked to evaluate sensory experiences, including a list of features that distinguishes their responses from those provided by non-synaesthetes.

In order to provide a fuller picture of the differences between the two groups in terms of the ways in which they responded to the prompts, it is useful to investigate their responses in a more qualitative manner. We consider here (a) the responses that were significantly likely to be favoured by the synaesthetes, (b) responses that were more frequent in the synaesthetes’ data but where the difference was not statistically significant and (c) the responses that were significantly more likely to be favoured by the non-synaesthetes. We now examine each of these response types in detail and explore specific examples. By bringing together these quantitative and qualitative findings, we hope to provide a picture of how the thinking patterns of synaesthetes differ from those of non-synaesthetes in terms of the factors that accompany the production of (creative) metaphor.

3.4 Conclusion

The findings from our analysis of the response types that tended to be favoured by the synaesthetes and non-synaesthetes respectively are that, in addition to providing significantly more metaphors, synaesthetes also made significantly more references to cognitive effects, emotions and physical effects, and personification. They also reported more extreme reactions to the experiences, as they were significant more likely to use ‘hyperbole’. They also employed somewhat more personification, empathy and references to what ‘should or should not be’. Their responses were also more elaborate. In contrast, the responses provided by the non-synaesthetes were more ‘straightforward’ and literal. The response types were also more correlated in the synaesthetes’ responses than in the non-synaesthetes’ responses. This suggests that there may be a cluster of behaviours that surround the production of creative cross-sensory metaphor that are particularly prevalent in synaesthetes, and that to some extent these behaviours may trigger the use of these metaphors.

Given these findings, we are now in a position to consider the overall differences between synaesthetes’ and non-synaesthetes’ evaluative descriptions of sensory experiences and what they might tell us about the relationship between emotional, physical and cognitive engagement and creative metaphor. One of the most notable observations from this study relates to the comparative ‘richness’ of the synaesthetes’ responses in comparison with the non-synaesthetes. We already saw in Section 2 that the synaesthetes had more to say than their non-synaesthete counterparts. In this section we have seen that not only do they say more, but they produce a wider variety of responses across the board in terms of response sense, type and metaphorical theme. They also appear to have less of a negativity bias than the non-synaesthetes.

We have also seen in this section that synaesthetes are more likely than non-synaesthetes to produce responses containing metaphor (most of which contained creative, cross-sensory mappings), and they were particularly likely to use metaphor when describing positive sensory experiences. They were also more likely to make reference to emotional, physical and cognitive effects, hyperbole and empathy. Finally, there were more likely to be inter-relationships between the responses types that were used by the synaesthetes. What do these co-occurrences mean and what might they say about the kinds of cross-sensory mappings that synaesthetes form, the factors that motivate them to make these mappings and the extent to which they can be described as ‘metaphorical’ or ‘creative’?

The co-occurrence between metaphor, emotion and cognitive effects in the synaesthetes’ data reflects findings in the metaphor literature that internal thought processes and experiences are often talked about using metaphor, as these are often difficult to express in literal terms (Reference Gibbs and FranksGibbs & Franks, 2002; Reference Littlemore and TurnerLittlemore & Turner, 2019; Reference Turner, Littlemore, Burgess, Barden and GargettTurner et al., 2020). Synaesthetes are more likely to report strong visceral responses to sensory experiences which can result in physical pain (‘it’s piercing through my skin’) or mental disturbances (‘they take over my head and drive me mad’), and they respond on a much more emotional level (‘it makes me anxious’). These extreme reactions appear to be linked to their tendency to form cross-sensory associations, many of which have a strong metaphorical feel to them (‘I like hearing cars zoom by because the sound looks like blue slowly fading’), and some of which involve personification (‘it’s indecisive about which pitch it wants to be’). All of these findings point to the embodied nature of creative metaphor generation, which was alluded to in Section 1. They suggest that the richer, and more intense, ways in which synaesthetes engage with sensory experiences may in fact be the trigger for many of the creative metaphors that they produce.

Metaphor has also been shown to be useful as an expressive tool, helping people to conceptualise and communicate experiences that are not widely shared by the rest of the population. With this in mind, it is particularly interesting to note that metaphor was relatively absent in the non-synaesthetes’ data. This may be because they are more likely to describe highly conventional emotional responses, which they do not feel the need to draw on metaphor to explain. Also, the experiences that the synaesthetes are describing tend to be more internal as they are often describing the thought processes themselves, whereas the non-synaesthetes tend to draw more on literal associations and memories of external experiences and actions. Compare the response from a synaesthete (a) with a response from a non-synaesthete (b):

The synaesthete in (a) draws on metaphor to describe the internal experience itself, whereas the non-synaesthete in (b) remembers a particular event without going into detail on the emotional quality of that memory. This means there is less scope for them to draw on metaphor in their response.

The correlation between cognitive effects and personification in the non-synaesthetes’ data may be due to the fact that personification allowed the non-synaesthetes to ‘blame’ particular sensory experiences for their responses, as in the following example:

These findings have interesting implications for the relationship between emotion, creativity and metaphor. It should be remembered that although the metaphorical mappings were creative, they all drew on motivated embodied relationships between physical and emotional experiences, which displayed varying degrees of conventionality. When a synaesthete comments that ‘those cookies tasted of regret and rotting flesh’ the synaesthete is extending an established embodied relationship between negative emotional experiences and bad flavours (e.g. ‘it left a bad taste in my mouth’) and juxtaposing this with a more perceptual metaphoric relationship. When a synaesthete comments that ‘Her voice is very heavy and thick and makes me feel like I’m sinking down slowly under dark water’, he or she is drawing on conventional, embodied metaphorical mappings between the target domain ‘sound’ and the source domains ‘weight’, ‘darkness’ and ‘movement’, and combining and extending them in a new way.

The reactions reported by the synaesthetes to different sounds were much more physical, emotional and intense than those reported by the non-synaesthetes, as we can see in the following example:

This may reflect findings from neuro-imaging studies where it has been shown that during sound perception, synaesthetes show significantly more brain activation in the inferior parietal cortex than non-synaesthetes (Reference Neufeld, Sinke and DilloNeufeld et al., 2012). This area is involved in multimodal integration, the perception of emotions in facial stimuli and, most importantly, the interpretation of sensory information.

To sum up, synaesthetic individuals reported stronger emotional, cognitive and physical effects in response to the sensory experiences and were somewhat more likely to personify the stimuli that triggered those experiences, empathise with, ‘judge’ them and describe the experiences in extreme terms. These highly ‘embodied’ responses to the experiences appear to have played a role in their production of creative metaphors, suggesting that embodied metaphor, empathy, emotion, physical experience and creativity are intrinsically linked through a symbiotic relationship, and that this link is particularly prominent in synaesthetes. Synaesthetes have a higher engagement with the sensory experiences, and a desire to communicate this to others. It has been shown in previous research that this kind of communicative need may shape the strategies that people employ when communicating rich sensory experiences (Reference WinterWinter, 2019), leading them to produce metaphor that can be considered to be creative (Reference Caballero, Forceville and Urios-AparisiCaballero, 2009). The responses produced by synaesthetes involved significantly more creative metaphor than those produced by the non-synaesthetes. The fact that synaesthetes’ responses were also more likely to involve empathy, references to physical, emotional and cognitive effects, ‘hyperbole’ and personification suggests that these factors may be implicated in the process of creative metaphor production. Overall, it appears to be the case that synaesthetes experience higher levels of emotional, cognitive and physical engagement with sensory experiences, which can lead them to produce vivid examples of metaphorical language and hyperbole as they explain them.

4.1 Introduction

In Section 3, we identified a number of differences in the ways in which synaesthetes and non-synaesthetes described sensory experiences. Our findings suggested that strong emotional engagement with a sensory experience may be a driver of metaphorical creativity. The fact that synaesthetes engaged with sensory experiences on a much more emotional level than the non-synaesthetes appears to have led them to empathise with these stimuli, experience physical and cognitive reactions to them, judge them as ‘right’ or ‘wrong’ and talk about their sensory experiences in creative metaphorical ways. If it is the case that heightened emotional experiences seem to lead synaesthetes to produce creative cross-sensory metaphors, is it also the case that they perceive emotion itself in different ways from non-synaesthetes? In order to answer this question, we explored synaesthetes’ and non-synaesthetes’ responses to emotions in more depth. We were interested in examining whether there are differences in the ways in which synaesthetes understand and respond to emotions. One of the best ways to access people’s understandings of a phenomenon is to ask them what they associate it with. By exploring the range of associations that people have with a given phenomenon, we gain in-depth knowledge of the way in which they understand and experience it.

In this section, we introduce a shorter test which participants undertook following the longer writing task. We used a word association task to investigate synaesthetes’ and non-synaesthetes’ responses to emotion-related words. They were asked to write the first thing that came to mind when they saw the word. As with the writing task discussed in Section 3, we did not specify the required length of the answers. The results provide insights into the different ways in which synaesthetes and non-synaesthetes understand and, by extension, experience different types of emotions. These insights will allow us to comment further on the relationships between creativity, metaphor and emotion in synaesthetic and non-synaesthetic populations, as first introduced in Section 3.

4.2 Investigating Responses to Emotion Words and Emotive Expressions

4.3 How Do Synaesthetes and Non-synaesthetes Differ in Their Responses to the Emotion Words WAT?

Here, we explore the responses to the word association task. We begin by observing the general differences between the two groups in terms of their responses to the emotion prompts in the word association task. The first notable difference relates to the overall number of words produced by the two groups, which mirrors the findings observed in our analysis of the sensory stimulus tasks. We observed in Section 2 that the body of writing produced by the synaesthetes in response to the sensory stimulus tasks was over twice as large as the dataset produced by the non-synaesthetes. We posited that synaesthetes’ sensory experiences may be somehow more rich or salient to them, although we noted that the self-selecting nature of the participants may also have played a role in this finding, as the synaesthetic participants in our study may have been particularly keen to make their experiences known. However, it is interesting to note that even in a word association task such as this one, the synaesthetes produced more text in response overall, with some synaesthetes providing extended descriptions of scenarios in response to the stimuli. Figures showing how the two groups differed in terms of their response word counts can be found in Table A12 in the additional resources. In this respect, the behaviours of each group in response to the word association task mirror the findings of the sensory description writing task. This finding may be taken as preliminary evidence that emotional stimuli provoke a similarly rich, vivid response in synaesthetes as do sensory experiences.

As well as producing an observable difference in word count, synaesthetes and non-synaesthetes differed in the extent to which they drew on the range of categories in their responses. The breakdown of response types according to their perceived conventionality is provided in Table 7. As mentioned in Section 4.2, responses could be coded as belonging to more than one category. These figures show that in comparison with the non-synaesthetes, synaesthetes produced more tokens overall (206 vs the non-synaesthetes’ 173), and they were significantly more likely to favour unconventional responses over conventional responses than the non-synaesthetes (χ² (1, N = 40) = 109.64, p >.00001). This mirrors the finding reported in Section 3 that synaesthetes produced more extensive and detailed descriptions of sensory experiences than non-synaesthetes. Similar patterns are therefore observed in synaesthetes’ responses to both sensory and emotional stimuli, which may provide further evidence for the highly embodied, emotionally rich nature of their experiences.

We can see from Table 7 that synaesthetes were more likely than non-synaesthetes to produce responses involving ‘actions’, ‘colour’, ‘nature and natural phenomena’, ‘people, body parts and facial expressions’, ‘physical objects’, ‘scenarios’, and ‘synaesthetic associations’. In contrast, non-synaesthetes were more likely than synaesthetes to produce responses involving ‘abstract concepts’, ‘synonyms and related emotions’, ‘exclamations and onomatopoeia’, ‘physical sensations’ and ‘qualities’. In order to establish which of these responses were significantly more likely to be favoured by one of the groups, we conducted Mann-Whitney U Tests. The results are shown in Table 8.

In order to explore the data in more detail, we also ran correlation analyses of the responses produced by the two groups to investigate whether particular types of responses correlated with each other, and whether these relationships held true across the two groups. We are aware that applying such statistical measures to small data sets such as ours is problematic, and that the results should be taken as indicative only, but we considered them a useful way of conceptualising the results and exploring patterns which would not otherwise be apparent. We therefore ran a series of Spearman correlation tests across all association types for each group of participants separately. The results are shown in Tables A13 and A14 in the additional resources.

The responses produced by the synaesthetes demonstrated significant collocations between: abstract concepts and synonyms/related emotions (ρ = .30); abstract concepts and nature/natural phenomena (ρ = .20); actions and people/body parts/facial expressions (ρ = .36); actions and scenarios (ρ = .58) and people/body/parts/facial expressions and scenarios (ρ = .50). There were significant negative correlations between colour and qualities (ρ = −.21), and colour and synaesthetic associations (ρ = −.33). The latter finding is particularly interesting as it suggests that when they used colour, the synaesthetes either provided a colour-based synaesthetic association, or they provided a metaphorical colour association, but they rarely combined both strategies. In contrast, there was only one significant correlation found in the non-synaesthetes’ data, in which a significant negative correlation was found between synonyms and related emotions and colour (ρ = −.20).

These analyses reveal that the responses produced by synaesthetes and non-synaesthetes differ in some notable ways. First, similarly to the writing task, the responses produced by the synaesthetes seem to demonstrate more interrelatedness in the categories they fall into. This is demonstrated by the fact that there are more significant correlations occurring in the synaesthetes’ responses. Perhaps the ‘interrelatedness’ in the categories here is reminiscent of the synaesthetes’ neural architecture, characterised as it is by highly levels of connectivity. It is interesting that the category denoting synaesthetic associations with letters and sounds does not correlate positively with any other response types. These associations are perhaps best considered their own separate category, a representation of a synaesthete’s own personal and arbitrary associations and perhaps, when triggered, strong enough to override the use of other response categories.

The statistical analyses reported above have enabled us to identify areas where the two groups noticeably differ from each other, and where they overlap. In order to illustrate these response patterns, it is useful at this point to explore the answers provided by some of the synaesthetes in a more qualitative manner. By doing so, we hope to show how the different categories of response related to one another, and how the divergent thinking patterns displayed by synaesthetes manifested when they were performing this task. We consider here (a) response types that were significantly more likely to be provided by synaesthetes, or only produced by them; (b) response types that were more frequent in the synaesthetes’ data but where the difference was not statistically significant; (c) response types that were favoured by non-synaesthetes, but where the difference was not statistically significant; and (d) response types that were significantly more likely to be provided by non-synaesthetes. Within each of these categories, we look at the way the two groups differ in their use of the response types in general, and the extent to which these response types were used creatively.

4.4 Conclusion

From our analysis of the associations that were favoured by the two groups, and of the qualitative differences between the ways in which the two groups used them, we can see that synaesthetes appear to favour responses that are more novel, more conceptually distant from the prompt, and that they engage with the prompt in a more personal or embodied way than the non-synaesthetes. They tend to provide longer, more detailed answers or extended scenarios that combine all of these features. In contrast, non-synaesthetes tend to favour more conventional associations that relate more closely and more directly to the stimulus. The synaesthetes’ response categories were also more likely to be strongly correlated with one another than those of the non-synaesthetes, demonstrating higher levels of interrelatedness. These findings speak to the associative thinking style that is characteristic of synaesthetes, which we discussed in Sections 1 and 3.

These findings suggest that synaesthetes engage with emotions in a somewhat different way to non-synaesthetes. They appear to engage with them on a deeper level and are more likely to relate them to personal physical experiences or interactions with the physical world. Their associations are much more likely to be unconventional. They combine novelty with a degree of appropriacy, which means that they are in a sense ‘creative’. In contrast, non-synaesthetes tend not to engage in cross-sensory or cross-domain mappings; they remain within the same conceptual domain, for example by associating one emotion with other related emotions. The ability to form wider sets of unconventional associations based on a broader set of criteria characterises the responses of the synaesthetes and appears to mirror earlier findings discussed in Section 1, which show that synaesthetes display higher levels of associative fluency. However again, like the responses discussed in Section 3, the associations that they form do not appear to be entirely arbitrary; rather, they involve novel elaborations of conventional associations. The tendency to produce these associations demonstrates higher levels of associative fluency in the synaesthete group, which has been shown to be a key factor underpinning metaphor.