Development of synaesthetic consistency: Repeated autonomous engagement with graphemes and colours leads to consistent associations
Introduction
Grapheme-colour synaesthesia describes consistent phenomenological experiences of colour in response to achromatic letters and/or numbers (Ward, 2013). For a synaesthetic individual, the letter A may always trigger a red colour experience (Baron-Cohen, Wyke, & Binnie, 1987). Although synaesthesia tends to run in families, pointing to a genetic basis (Asher et al., 2009), synaesthetic inducers are often cultural artefacts, suggesting a learning component in its development (Simner and Bain, 2013, Simner et al., 2009). For example, synaesthetic associations may reflect environmental contingencies that were consolidated through repeated engagement with grapheme-colour associations during childhood (Witthoft and Winawer, 2006, Witthoft et al., 2015). Although plausible, it remains to be investigated if repeated engagement with unspecified grapheme-colour associations is sufficient to result in consistent grapheme-colour associations (i.e., the hallmark of grapheme-colour synaesthesia). Thus, we tested in non-synaesthetes whether repeatedly engaging with grapheme-colour associations promotes synaesthetic consistency.
There is general agreement that synaesthetic associations are consistent over time and tests of consistency have become the gold standard to confirm the genuineness of any type of synaesthesia (Baron-Cohen et al., 1987, Carmichael et al., 2015, Eagleman et al., 2007, Rothen et al., 2013). During a consistency test, participants are repeatedly asked to indicate their synaesthetic experiences for potential synaesthetic inducers. While synaesthetes can rely on their synaesthesia when performing such a test, non-synaesthetes are dependent on their memory. Therefore, synaesthetes generally outperform non-synaesthetes in consistency tests (e.g., Rothen et al., 2013).
To investigate the development of grapheme-colour synaesthesia, Simner et al. (2009) conducted a consistency test with more than 600 children at the age of 6–7 years, including a re-test 12 months later. Moreover, a subsample of 80 children (synaesthetes and non-synaesthetes) was tested again at the age of 10–11 years (Simner & Bain, 2013). The consistency test was based on a palette of 13 distinct colours and consistency was measured as percentage of identical grapheme-colour associations between different runs or sessions. Five of eight synaesthetes identified in the original study were still classified as synaesthetes at the age of 10–11 years. Crucially for the purpose of our study, these studies showed an increase in the percentage of consistent synaesthetic associations (henceforth synaesthetic consistency) from an average of 34% at age 6–7 years to 48% at 7–8 years and to 71% at 10–11 years. The data suggest that synaesthetic consistency, at least for grapheme-colour synaesthesia, seems to gradually develop in childhood during the first school years when children learn to read and write.
Related to the notion that synaesthesia develops during the first school years, Watson et al., 2014, Watson et al., 2017) proposed that grapheme-colour synaesthesia develops as a consequence of learning strategies, which are adopted when learning to deal with the meaning of abstract information of complex systems of cultural artefacts (e.g., reading and writing). Recently, Watson et al. (2017) provided empirical evidence that learning challenges in childhood may promote the development of grapheme-colour synaesthesia. They tested whether growing up with an orthographically more difficult language (English) leads to an enhanced prevalence of synaesthesia in contrast to growing up with an orthographically easier language (Czech). Although the main hypothesis was not confirmed, further analyses revealed the expected relation between orthographic difficulty and prevalence of synaesthesia when children growing up with only one language (apart from Czech) in the first two years of life were compared. The authors speculated that children growing up with a more difficult language build synaesthetic associations to support learning their mother tongue. Crucially, such a strategy would entail repeated engagement with unspecified grapheme-colour associations which become gradually consolidated over time.
Others have suggested that synaesthetic associations are shaped by environmental contingencies (Hancock, 2006, Rich et al., 2005, Witthoft and Winawer, 2006, Witthoft and Winawer, 2013, Witthoft et al., 2015). For instance, a relation between coloured letter-toys and synaesthetic associations has been found not only in case studies (Hancock, 2006, Witthoft and Winawer, 2006) but also in larger samples (Rich et al., 2005, Witthoft and Winawer, 2013, Witthoft et al., 2015). More recently, Witthoft et al. (2015) showed that for 400 out of 6588 grapheme-colour synaesthetes, their synaesthetic associations could be traced back to one specific coloured magnetic-letter set. Witthoft and Winawer (2013) proposed that grapheme-colour synaesthesia can develop as a consequence of conditioned mental imagery. That aside, the data suggest at least that repeated engagement with coloured letters may result in consistent grapheme-colour associations which may further develop into synaesthetic phenomenology.1
Furthermore, non-random patterns of grapheme-colour associations in large samples of grapheme-colour synaesthetes and non-synaesthetes (Simner et al., 2005) may be interpreted as indirect evidence for an influence of environmental contingencies during the development of synaesthesia (cf. also Mankin and Simner, 2017, Root et al., 2018) and cross-modal correspondences in non-synaesthetes, which have previously been interpreted as a weak form of synaesthesia (Martino & Marks, 2001). Simner et al. (2005) compared grapheme-colour associations in consistency tests from over 300 non-synaesthetes with the associations of synaesthetes. The results showed that both groups tended to associate some graphemes with certain colours more often than would be expected by chance (e.g., A-red). Moreover, there was also substantial overlap between the specific grapheme-colour associations of the two groups, suggesting a common mechanism behind the development of synaesthetic consistency and cross-modal correspondences. Specifically, synaesthetic associations might at least be partly determined by consolidated cross-modal correspondences (cf. also Newell & Mitchell, 2016).
Based on our brief literature review, it seems plausible to assume that the acquisition of initially unspecified grapheme-colour associations is based on a consolidation process of environmental contingencies, possibly due to self-employed learning strategies when mastering the skills of reading and writing. The assumption of a consolidation process finds further support in frequent anecdotal reports of synaesthetic siblings who discussed with each other during childhood whose grapheme-colour associations are “correct” (i.e., consolidation due to repeated engagement with grapheme-colour associations). Since also non-synaesthetes are subjected to the same environmental contingencies as synaesthetes (Mankin & Simner, 2017), one can reasonably expect them to exhibit the hypothesized consolidation mechanism when they are required to repeatedly engage in autonomously associating colours to letters and to develop consistent grapheme-colour associations (henceforth, grapheme-colour consistency).
To test this hypothesis, we had two groups of non-synaesthetes engage in a daily grapheme-colour association task (i.e., a simplified version of a categorical synaesthetic consistency test) over a period of ten days, without any reference to synaesthesia (henceforth grapheme-colour engagement phase). Participants of both groups were instructed to select colours for each letter of the alphabet and the digits 0–9, based on their gut feeling. One group was further instructed to memorize the selected colour for each of the graphemes. The other group received no further instructions. Crucially, grapheme-colour associations were not specified, by contrast to synaesthesia training studies with predefined grapheme-colour associations (e.g., Bor et al., 2014, Colizoli et al., 2014, Rothen et al., 2018). Both groups completed a continuous consistency test in a session before (henceforth pre-session) and after (henceforth post-session) the engagement phase. A third group completed only the pre- and post-session, with an unfilled interval of ten days in between. We expected an asymptotic increase in grapheme-colour consistency in both experimental groups over the course of the engagement phase. Moreover, we expected a faster increase in grapheme-colour consistency in the group that was required to memorize the self-selected grapheme-colour associations in comparison to the group without this requirement. We expected little to no increase in grapheme-colour consistency for the group that completed the consistency test before and after an unfilled interval.
Furthermore, all three groups completed two recognition memory tests (i.e., words and fractals) in the pre- and post-session: first, to confirm the groups did not differ with regards to memory performance before the engagement phase; second, we did not want to miss the opportunity to explore if the potential acquisition of synaesthetic associations enhances memory performance, as is the case in genuine synaesthesia (cf. Rothen, Meier, & Ward, 2012). However, we suspected that the engagement phase might be too short to affect memory performance.
Section snippets
Participants
We tested a total of 74 participants, who were randomly assigned to one of three groups: a group that had to select colours for letters on basis of their gut feeling (i.e., implicit group), a group that was additionally instructed to memorize the selected colours (i.e., explicit group), and a group with an unfilled interval (i.e., control group). Inclusion criteria were normal (or corrected to normal) vision and absence of synaesthesia (based on phenomenological self-reports). Three
Continuous consistency test
The continuous consistency test was used to assess baseline performance in the pre-session and a potential transfer of benefits in the post-session, due to the engagement phase with discrete colours. Descriptive statistics for the continuous consistency test are shown in Fig. 3a as a function of Group (explicit, implicit, control) and Session (pre vs. post). For comparison, data of a synaesthetic sample are provided in Fig. 3a. A two-factorial ANOVA for mixed measures with the between subject
Discussion
Synaesthetic consistency is the hallmark of synaesthesia and plays an important role in the definition and validation of synaesthesia. However, the exact nature of its development is still a matter of debate. The main finding of the present study is that repeated engagement in autonomously associating graphemes with colours can lead to highly consistent grapheme-colour associations in non-synaesthetes. An increase in grapheme-colour consistency was also observed when participants were not
Conclusion
In summary, we established consolidation processes based on repeated engagement with unspecified synaesthetic inducer-concurrent pairings as a potential mechanism in the development of synaesthetic consistency. Crucially, the observed pattern of the developmental trajectory in our non-synaesthetic adult sample closely mimicked the pattern of the developmental trajectory in genuine synaesthesia. Interestingly, the acquisitions of consistent grapheme-colour associations through engagement with
Declaration of Competing Interest
None.
Acknowledgements
ROF and NR are supported by the Swiss National Science Foundation (Grant Number PZ00P1_154954). We thank Beat Meier for helpful methodical discussions. We also thank Eleonora Balbi for helpful comments on an earlier version of the manuscript, and Sophie Ankner, Esther Brill, Sarah Di Pietro and Christopher Ritter for their help with data acquisition.
References (48)
- et al.
A whole-genome scan and fine-mapping linkage study of auditory-visual synesthesia reveals evidence of linkage to chromosomes 2q24, 5q33, 6p12, and 12p12
American Journal of Human Genetics
(2009) - et al.
Validating a standardised test battery for synesthesia: Does the Synesthesia Battery reliably detect synesthesia?
Consciousness and Cognition
(2015) - et al.
A standardized test battery for the study of synesthesia
Journal of Neuroscience Methods
(2007) Monozygotic twins’ colour-number association: A case study
Cortex
(2006)- et al.
Multisensory integration and cross-modal learning in synaesthesia: A unifying model
Neuropsychologia
(2016) - et al.
Enhanced associative memory for colour (but not shape or location) in synaesthesia
Cognition
(2013) - et al.
A systematic, large-scale study of synaesthesia: Implications for the role of early experience in lexical-colour associations
Cognition
(2005) - et al.
Why is the synesthete’s “A” red? Using a five-language dataset to disentangle the effects of shape, sound, semantics, and ordinality on inducer–concurrent relationships in grapheme-color synesthesia
Cortex
(2018) - et al.
Enhanced memory ability: Insights from synaesthesia
Neuroscience and Biobehavioral Reviews
(2012) - et al.
Coordinated neural, behavioral, and phenomenological changes in perceptual plasticity through overtraining of synesthetic associations
Neuropsychologia
(2018)