Explorations of Sound Symbolism and Iconicity

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2019-09-17 Explorations of Sound Symbolism and Iconicity

Sidhu, David Michael

Sidhu, D. M. (2019). Explorations of Sound Symbolism and Iconicity (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/111016 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca

UNIVERSITY OF CALGARY

Explorations of Sound Symbolism and Iconicity

David Michael Sidhu

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN PSYCHOLOGY

CALGARY, ALBERTA

SEPTEMBER, 2019

Abstract

Sound symbolism refers to the finding that individuals have biases to associate certain language sounds (i.e., phonemes) with certain perceptual and/or semantic features (see

Lockwood & Dingemanse, 2015; Sidhu & Pexman, 2018a). An example of this is the association between the phoneme /i/ (as in heed) and smallness. This is of special interest to language because it can enable iconic relationships between form and meaning: instances in which a word’s form maps onto its meaning via resemblance. For instance, the word teeny contains a vowel associated with smallness, and refers to something small. Iconicity can also exist through direct resemblance, in which a form imitates the meaning to which it refers (e.g., bang, woosh).

In Chapter 2 I synthesize the existing sound symbolism literature to arrive at five potential mechanisms that could give rise to the associations between phonemes and features. I also discuss as yet unanswered questions for the field and propose ways in which future research might answer these questions.

In Chapter 3 I demonstrate a novel form of sound symbolism, namely that between phonemes and personality factors. In a departure from much of the previous literature, I conduct this investigation using real first names, allowing exploration of sound symbolism in existing language. Further, by demonstrating an association between phonemes and an abstract dimension, I widen the scope of sound symbolism, and provide a novel test case for the potential mechanisms discussed in Chapter 2.

In Chapter 4 I turn my attention to iconicity and its benefit to language processing. I demonstrate that iconic words are processed faster on a lexical decision task as well as a phonological lexical decision task, compared to arbitrary words. I consider how this finding might fit into an existing model of word processing.

Finally, in Chapter 5, I explore the effect of iconicity on the structure of the lexicon. I demonstrate that iconic words tend to have more unique meanings, and to have a greater amount of associated sensory experience. I discuss how these findings could shed light on the emergence of iconicity in the lexicon over time.

Across these diverse studies I explore non-arbitrariness in language both at the level of individual phonemes and entire words. A running theme throughout this work is a consideration of the mechanisms underlying these phenomena, as well as an exploration of their relevance to broad, existing language.

iii

Preface

Chapter 2 is adapted from the pre-proof version of the following open access publication, permitted under the terms of the Creative Commons Attribution 4.0 International License

(http://creativecommons.org/licenses/by/4.0/):

Sidhu, D. M., & Pexman, P. M. (2018). Five mechanisms of sound symbolic association. Psychonomic Bulletin & Review, 25, 1619-1643.

Sidhu, D. M., Deschamps, K., Bourdage, J. S., & Pexman, P. M. (2019). Does the name say it all? Investigating phoneme-personality sound symbolism in first names. Journal of Experimental

Psychology: General. doi:10.1037/xge0000662

This article may not exactly replicate the authoritative document published in the APA journal. It is not the copy of record. No further reproduction or distribution is permitted without written permission from the American Psychological Association.

Chapter 4 is a pre-proof version of the following manuscript, which has been accepted for publication:

Sidhu, D. M., Vigliocco, V., & Pexman, P. M. (in press). Effects of iconicity in lexical decision.

Language and Cognition.

Sidhu, D. M., & Pexman, P. M. (2018). Lonely sensational icons: semantic neighbourhood density, sensory experience and iconicity. Language, Cognition and Neuroscience, 33, 25-31.

Acknowledgements

Thank you Penny Pexman for being the best supervisor anyone could ask for. If I were to rerun the simulation of grad school 100 times, I would hope to be lucky enough to be your student each of those 100 times. None of this would have happened without your wisdom, support and guidance.

Thank you Mandy for your love, companionship and understanding. You’ve been a wonderful partner throughout this process and I can’t imagine having done it without you. I will always be grateful for the times you reassured me when things went wrong, and the times you celebrated with me when things went right. {~}

Thank you Mom for fostering my imagination, and your unshakable love and belief in me.

Thank you Dad for instilling in me the value of hard work and of not giving up.

Thank you Puba, Jida, Kamar, Daniel, Guy, Felix, Noah, Adev, Tracy, Gord and all of my other family for your constant support over the years.

Thank you to my friends for making life a joy.

Thank you to all the lab members, graduate students and professors who helped me ask, and look for answers to, interesting questions.

Thank you to my committee members Glen Bodner, Suzanne Curtin, Darin Flynn and Jamie

Reilly for your time and thoughts.

Table of Contents

Chapter 1: Introduction 1 Sound Symbolism 1 Arbitrariness, Systematicity and Iconicity 4 The Studies to Follow 9 Chapter 2 9 Chapter 3 9 Chapter 4 10 Chapter 5 10

Chapter 2: Five Mechanisms of Sound Symbolic Association 12 Sound Symbolism 12 Size and Shape Symbolism 16 Arbitrariness and Non-Arbitrariness 20 Phonetic Features Involved in Sound Symbolism 27 Mechanisms for Associations Between Phonetic and Semantic Features 30 Mechanism 1: Statistical Co-Occurrence 32 Mechanism 2: Shared Properties 36 Low-level properties 37 High-level properties 38 Mechanism 3: Neural Factors 46 Mechanism 4: Species-General Associations 48 Mechanism 5: Language Patterns 51 Contextual Factors 53 Outstanding Issues and Future Directions 55 Phonetic Features 55 Relationship with Crossmodal Correspondences 58 Next Steps in Exploring Mechanisms of Association 60 Conclusion 65 Acknowledgements 68

Chapter 3: Investigating Phoneme-Personality Sound Symbolism in First Names 69 Introduction 69 Sound Symbolism 69 Sound Symbolism in Real Language 73 Phoneme-Personality Sound Symbolism 74 The Present Study 80 Experiment 1 83 Method 83 Ethics Statement 83 Participants 83 Materials and Procedure 84 Results 86 Discussion 89 Experiment 2 90 Method 90

vii

Participants 90 Materials and Procedure 90 Results 91 Discussion 93 Experiment 3 93 Method 93 Participants 93 Materials and Procedure 94 Results 95 Discussion 105 Experiment 4 106 Method 106 Participants 106 Materials and Procedure 106 Results 108 Discussion 110 Experiment 5 111 Method 111 Participants 111 Materials and Procedure 112 Results 112 Discussion 114 General Discussion 115 Mechanisms for Phoneme-Personality Sound Symbolism 119 Real World Implications 124 Conclusion 126 Acknowledgements 127

Chapter 4: Effects of Iconicity in Lexical Decision 128 Introduction 128 Experiment 1 132 Methods 132 Participants 132 Materials and Procedure 132 Results 135 Statistics 135 Reaction time 136 Accuracy 137 Discussion 139 Experiment 2 139 Methods 139 Participants 139 Materials and Procedure 139 Results 140 Reaction time 140 Accuracy 142 Discussion 144 English Lexicon Project Analysis 144 Results 144 Discussion 146

viii

General Discussion 147 Conclusion 150 Acknowledgements 151

Chapter 5: Semantic Neighbourhood Density, Sensory Experience and Iconicity 152 Introduction 152 Method 156 Materials and Procedure 156 Results 157 Discussion 163 Acknowledgements 166

Chapter 6: Conclusion 167 Limitations 168 Future Research 170 Sound Symbolism 170 Iconicity 171 Theoretical Contributions 172 Future Theoretical Refinements 174 Sound Symbolism 174 Iconicity 175

References 177

Appendix A 209

Appendix B 211

Appendix C 213

List of Tables

Table 1 Sample definitions of phonetic sound symbolism in the literature. 13 Table 2 Definitions of linguistic terms used throughout the paper (derived from Ladefoged & Johnson, 2010; Reetz & Jongman, 2009). 18 Table 3 A summary of the shared properties that are proposed to be involved in sound symbolism. 45 Table 4 Definitions of linguistic terms used throughout the article (derived from Ladefoged & Johnson, 2010; Reetz & Jongman, 2009). 71 Table 5 “Round” and “Sharp” traits generated by participants and used as stimuli by Sidhu and Pexman (2015) in their Experiment 2. 79 Table 6 A description of each factor of the HEXACO, taken verbatim from Lee and Ashton (2009). 81 Table 7 Resulting intercepts of logistic regressions predicting the likelihood of selecting a sonorant (voiceless stop) name for traits from the high (low) end of each personality factor, in Experiment 1. 88 Table 8 Resulting name type coefficients of linear regressions predicting fit ratings of sonorant (voiceless stop) names for traits from the high (low) end of each personality factor, in Experiment 2. 92 Table 9 Zero-order correlations among predictor and outcome variables in Experiment 3. Correlations between the adjective- and statement-based measures of a personality factor are underlined. 96 Table 10 Mean, Standard Deviation (calculated across subjects) and reliability for each of the personality inventories used in Experiment 3. 97 Table 11 Resulting coefficients of multivariate linear regressions predicting participant scores for each personality factor as a function of name phonology, in Experiment 3. 100 Table 12 Resulting coefficients of multivariate linear regressions predicting participant scores for each personality factor as a function of nickname phonology, in Experiment 3. 103 Table 13 Resulting name type coefficients of linear regressions predicting fit ratings of sonorant (voiceless stop) invented names for traits from the high (low) end of each personality factor, in Experiment 4. 108 Table 14 Results of mediational analysis in Experiment 4. Analysis was performed at the item level, with name type as the independent variable, and perceived name gender as a mediator variable. 110 Table 15 Results of mediational analyses in Experiments 2 and 4. Analyses were performed at the item level, with name type as the independent variable, and name likeability as a mediator variable. 114 Table 16 Potential mechanisms underlying sound symbolic associations from Sidhu and Pexman (2018). 120 Table 17 Mean (SD) values of lexical and semantic variables for each of the word types presented in Experiments 1 and 2. 134 Table 18 Linear mixed effects regression model predicting LDT reaction time in Experiment 1. 137 Table 19 Logistic mixed effects regression model predicting LDT accuracy in Experiment 1. 138 Table 20 Linear mixed effects regression model predicting PLDT reaction time in Experiment 2. 141 Table 21 Logistic mixed effects regression model predicting PLDT accuracy in Experiment 2. 143 Table 22 Correlations among variables. 157 Table 23 Results of hierarchical regression predicting iconicity. 159 Table 24 Correlations between iconicity and ARC and SER, for each word type. 160

List of Figures

Figure 1. The percentage of psychological publications per year that included the term “sound symbolism” and/or “iconicity”, according to PsycINFO. 16 Figure 2. The sign for bird in American Sign Language. 23 Figure 3. An illustration of vowel space. 29 Figure 4. An illustration of the different ways in which denotative and/or connotative dimensions may result in mediated relationships. 42 Figure 5. The relationship between sound symbolic associations and crossmodal correspondences (as the terms are used in this review). 59 Figure 6. An illustration of the maluma/takete effect, in which most individuals judge the nonword maluma as a better match for the round shape on the left, and the nonword takete as a better match for the sharp shape on the right. 70 Figure 7. The results of Experiments 1, 2 and 4, for each personality factor. 116 Figure 8. Plot showing the relationship between iconicity and predicted reaction time in the non-degraded LDT from Experiment 1. 146 Figure 9. The moderating effect of ARC on SER’s relationship with iconicity. 161 Figure 10. Scatterplots show the relationships between SER/ARC and iconicity, separately for adjectives/adverbs, verbs and nouns. 162

Epigraph

There are two ways to escape suffering [the inferno]. The first is easy for many: accept the inferno and become such a part of it that you can no longer see it. The second is risky and demands constant vigilance and apprehension: seek and learn to recognize who and what, in the midst of the inferno, are not inferno, then make them endure, give them space.

–Italo Calvino, Invisible Cities

Believed to have hung on David Sidhu’s office wall throughout graduate school.

Chapter 1: Introduction

A spoken language consists of an inventory of sounds (i.e., phonemes). These sounds are combined to form words. Words are then associated with meaning(s). There are two fundamental questions relating to this process that form the bedrock of this thesis. The first relates to phonemes: do phonemes themselves have some inherent association(s) with meanings? That is, are phonemes semantically neutral building blocks, or do their individual qualities imbue them with some semantic associations. The phenomenon of sound symbolism (see below) suggests that they do carry associations. How these associations are formed will be the focus of Chapters 2 and 3. The second question pertains to words: do the forms of words have any special relationships with their meanings or are the two connected in a purely arbitrary manner?

Chapters 4 and 5 will focus on the phenomenon of iconicity (see below), and the effect that this form of non-arbitrariness has on language processing and on the makeup of the lexicon in general.

Sound Symbolism

Sound symbolism is defined in slightly different ways by different authors. The way I conceptualize the phenomenon is that certain individual phonemes (or phoneme features)1 have associations with certain perceptual and/or semantic properties (for similar definitions see

D’Onofrio, 2013; Hirata, Ukita, & Kita, 2011; Koriat & Levy, 1977). Thus, for instance, hearing

(or reading, and generating the phonological code for) the phoneme /i/ evokes the property of smallness. The word “evoke” is admittedly vague. Accepting that this might not be entirely accurate, one way of operationalizing it would be to say that the phoneme /i/ activates the

1 There is some uncertainty as to whether sound symbolism exists for a phoneme as a whole (i.e., /i/ in its entirety), or rather a certain phoneme feature (e.g., a high second formant). See Monaghan and Fletcher (2019), and Westbury, Hollis, Sidhu, and Pexman (2018) for recent attempts at distinguishing the two possibilities.

2 perceptual/semantic feature of smallness. Importantly, this association and resulting activation is a result of properties of the phoneme itself (i.e., rather than being due to other words containing this phoneme). Something about /i/ itself leads to an association with smallness. Chapter 2 will explore the possible mechanisms underlying such associations.

The association between /i/ and smallness is an example of the mil/mal effect. It consists of an association between high-front vowels (e.g., /i/ as in heed) and small shapes, and low-back vowels (e.g., /ɑ/ as in hawed) and large shapes (Newman, 1933; Sapir, 1929). Note that Table 1 in Chapter 2 contains a description of linguistic terms used throughout. This demonstrates that sound symbolic associations typically consist of groups of phonemes sharing some phonetic feature being associated with a particular perceptual and/or semantic feature. The other most well-known example of sound symbolism is the maluma/takete effect (Köhler, 1929). It refers to an association between certain phonemes and either visual roundness or sharpness. Though the maluma/takete effect is nearly 100 years old, there is not a consensus on which phonemes are involved. McCormick, Kim, List and Nygaard (2015) made an important contribution to the field in a series of studies that had participants judge the fit between various nonwords, and round and/or sharp shapes. The picture that emerged was that sonorants (e.g., /m/, /n/, /l/) were especially associated with round shapes while voiceless stops (e.g., /p/, /t/, /k/) were especially associated with sharp shapes. Voiced stops (e.g., /b/, /d/, /g/) were somewhat associated with round shapes. Voiceless fricatives (e.g., /f/, /s/) were somewhat associated with round shapes while voiced fricatives (e.g., /v/, /z/) were somewhat associated with sharp shapes. This suggests that voicing, place of articulation and manner of articulation might all be relevant to the maluma/takete effect. Moreover, these different features might interact–note that the effect of voicing was opposite for stops and fricatives.

It is even less clear which vowels are implicated in the maluma/takete effect. In general studies have considered back vowels (e.g., /u/ as in who’d, /ɑ/) to be associated with roundness, and front vowels (e.g., /i/) to be associated with sharpness. Because frontness is almost perfectly confounded with roundness in English, this has also meant that round-associated vowels have been rounded, while sharp-associated vowels have been unrounded. The association between back/rounded vowels and roundness, and front/unrounded vowels and sharpness was confirmed by McCormick et al. (2015). At present it is unclear which of these two features contributes more to the effect.

The maluma/takete effect has been demonstrated in several different ways. The most common method is a forced choice task in which individuals are asked which nonwords go along with which shapes (e.g., Nielsen & Rendall, 2011). Participants are given a nonword containing round-associated phonemes, and a nonword containing sharp-associated phonemes, and then are asked to pair them with a round and sharp shape. Roughly 90% of individuals make the congruent pairing (see Styles & Gawne, 2017). Some have gone beyond this forced choice method to include a rating scale of the fit between nonwords and shapes (e.g., Cuskey, Simner,

& Kirby, 2015). Other demonstrations have included an implicit association task (Parise &

Spence, 2012), a priming task (Sidhu & Pexman, 2016) and a task involving continuous flash suppression (Hung, Styles, & Hsieh, 2017). There has also been neuroimaging evidence of different responses to nonword/shape pairings that are either congruent or incongruent (e.g.,

Asano et al., 2015).

While the majority of studies have used English-speaking undergraduates, some have examined the universality of sound symbolism. The maluma/takete effect appears to be reliably present by one year of age (e.g., Asano et al., 2015; Pejovic & Molnar, 2016; for a review see

Fort et al., 2018). The effect has also been demonstrated in a variety of cultures and language groups, including the Himba people of Namibia, a semi-nomadic group speaking a language without a written form (Bremner et al., 2013; for a review see Styles & Gawne, 2017). There are two notable failures to replicate the effect, in speakers of a Hunjara dialect (Rogers & Ross,

1975) and Syuba (Styles & Gawne, 2017). Styles and Gawne (2017) speculated that this was related to the phonetic/phonotactic legality of stimulus words in these languages. The effect also seems to be negatively correlated with Autism Quotient (Gold & Segal, 2017).

In addition to shape and size, sound symbolism has also been demonstrated for features such as: brightness (e.g., Newman, 1933), speed (Cuskley, 2013), hue (Moos et al., 2014), taste

(Gallace, Boschin, & Spence, 2011), and social dominance (Auracher, 2017). There was also work in the middle 20th century exploring different associations between phonemes and semantic differential scales of the sort used by Osgood, Suci, and Tannenbaum (1963; e.g., pleasant- unpleasant, active-passive; Greenberg & Jenkins, 1966; Miron, 1961). In Chapter 3 I will describe a novel association between phonemes and personality.

Arbitrariness, Systematicity and Iconicity

Sound symbolism refers to associations between phonemes and certain perceptual/semantic features. I will now turn my attention to what happens when these phonemes are packaged together into words and these words become associated with meanings. In particular, what is the nature of the relationship between words and their meanings. As will be discussed, the sound symbolism of a word’s component phonemes plays a role in that relationship, but that is not the only consideration.

To preface this discussion, it is worth first considering the different ways in which something can refer to something else. In semiotic terms, this is the relationship between a

5 signifier (i.e., something referring to something else) and its signified (i.e., the thing to which it refers). Charles Saunders Peirce (1974) set out three possible categories for this relationship. The first is a wholly symbolic relationship. In this case, there is no pre-existing or special relationship between the signifier and signified. Instead, a group of people agree that one refers to the other, and so the relationship is created and learned. Thus, a symbolic relationship is a conventional, arbitrary and person-made one. One must learn that a red octagon signifies the need to stop–there is no natural relationship between them. In an indexical relationship the signifier exists as a result of the signified, and thus refers to it by way of its very existence. A common example is smoke signifying fire. Finally, in an iconic relationship, the signifier refers to the signified by resembling it in some way. A road sign with a picture of traffic lights signifies that traffic lights are ahead by physically resembling them.

These categories have framed the discussion of the nature of the relationship between words and their referents. In particular, we are concerned with the forms of words (i.e., their sound, visual appearance, and associated articulatory sensations)–in particular their phonological forms (i.e., their sound). In a recent review, Dingemanse, Blasi, Lupyan, Christiansen, and

Monaghan (2015) laid out the three ways in which words could be related to their referents: arbitrarily, systematically and iconically. These will be reviewed here in brief, but are explored in more depth in Chapter 2. Arbitrariness recalls Peirce’s symbolic relationship. In an arbitrary relationship, there is no special link between the sound of a word and its meaning. The two are connected purely by convention, and this pairing needs to be learned (i.e., it cannot be deduced from properties of the word itself). One could likely open the dictionary at random and find examples of arbitrariness. There is no reason why the sounds in cat should refer to a cat; the

6 relationship is largely2 arbitrary–a product of convention produced through accidents of history.

There is nothing about these sounds that make cat a better match for its meaning than the sounds in dog or bird. We have simply learned a relationship between cat and its meaning.

Arbitrariness has been called a fundamental property of language (e.g., Hockett, 1963).

Indeed, it was one of the properties of language that Charles Hockett believed set it apart from animal communication. Arbitrariness allows human language to refer to any possible concept, by not requiring a special relationship between the two. At this point it is helpful to make a subtle distinction. While language may not require non-arbitrariness, this does not necessarily mean that it does not contain non-arbitrariness. A person does not require shoes to walk, but this does not mean that people do not walk with shoes on. The last few decades have seen a growing appreciation of the fact that language also contains non-arbitrary elements.

One example of non-arbitrariness is systematicity. It refers to large-scale patterns in the forms of words belonging to the same category. For instance, there are differences in the forms of nouns and verbs (e.g., in their stress pattern; Sereno, 1986). There are also differences in the forms of abstract and concrete nouns (e.g., abstract nouns tend to have more syllables; see Reilly

& Kean, 2007). Systematicity represents a kind of non-arbitrariness because a word’s form has some relationship with its meaning. However, systematicity doesn’t easily fit into the categories of indexicality or iconicity. This is because Peirce’s categories pertained to specific words and meanings, while systematicity refers to broad patterns, across many words rather than a specific one.

Another example of non-arbitrariness is iconicity: instances in which the form of a word resembles its meaning in some way. Theorists tend to talk about iconic “mappings” (e.g.,

2 To circumvent the counter-argument that few words are wholly arbitrary, which will become apparent in a few paragraphs, and at the latest by Chapter 2.

Emmorey, 2014), in which elements of form map onto elements of meaning. This brings up two key points: 1) an iconic resemblance need not be perfect, and 2) the mapping can exist between parts rather than the whole. To illustrate these two points, consider the word cricket, which resembles the sound of the insect to which it refers. Notice that it does not perfectly resemble its sound, it is rather an approximation. In addition, the word’s sound maps onto part of the word’s meaning (i.e., the sound that it makes).

Iconic mappings can occur in several ways. The most obvious is direct iconicity (see

Masuda, 2007), in which the sound of a word maps onto its meaning directly. Since this involves a straightforward resemblance, direct iconicity necessarily involves words with auditory meanings. It would be impossible for a phonological form to directly resemble a visual feature.

English examples of direct iconicity include words for: animal sounds (e.g., meow, quack, moo), other sounds (e.g., thunder, bang, buzz), and actions with prominently associated sounds (e.g., smooch, sizzle, splash). These words are known as onomatopoeia and they “appear consistently across all spoken languages” (Perniss, Thompson, & Vigliocco, 2010, p. 2).

Iconic mappings can also occur indirectly, in what I will refer to as indirect iconicity (see

Masuda, 2007). In this kind of iconicity, a word maps onto a meaning indirectly, by way of something else. While a phonological form couldn’t directly resemble a visual feature, it might be evocative of certain visual features and so map onto visual features in this indirect way. This is the relevance of sounds symbolism to iconicity. The phonemes of a word might evoke certain features by way of a sound symbolic association, allowing for an iconic mapping. Consider the word mini. The phoneme /i/ is associated with smallness, and the word itself refers to smallness.

We might consider this an instance of indirect iconicity, in which a word’s form maps onto its meaning indirectly via the associations it conjures.3

It is worth pausing to distinguish the terms sound symbolism and iconicity, because they are sometimes used interchangeably in the literature. In this thesis, I use the term iconicity to refer to a type of relationship between the form of a word and its meaning. Specifically, a relationship in which form resembles meaning. I use the term sound symbolism to refer to a relationship between a certain phoneme (and/or phoneme feature), and perceptual and/or semantic features. The association between /i/ and smallness is an example of sound symbolism.

Sound symbolism can enable iconicity. When a word that contains /i/ refers to something small, that is an example of iconicity.

Importantly, words are not expected to fall wholly into the categories of arbitrary, systematic or iconic. Instead, a word contains each of these properties in varying amounts

(Dingemanse et al., 2015). This is to be expected, as any word is a bundle of phonological, orthographic and articulatory features, in addition to its suprasegmental properties (e.g., stress pattern), and the distributional properties of its features in the lexicon as a whole (e.g., the frequency of its bigrams). These various features, as well as different instances of each feature

(e.g., the different phonemes in a word) could have different relationships with its meaning. To further complicate matters, a word’s meaning is not unidimensional. Instead, it is likely to be made up of a variety of semantic dimensions, each representing different types of semantic information (e.g., Pexman, Siakaluk, & Yap, 2014); it also belongs to various categories within a

3 Some suggest that indirect iconicity cannot be understood in a single form, but that it makes use of what Peirce (1974) termed diagrammatic iconicity (see Dingemanse et al., 2015). That is, a relationship between two forms resembling the relationship between two meanings. For instance, /i/ might only seem small in relation to another vowel (e.g., /ɑ/), which resembles the relationship between a word like mini and words referring to larger things.

9 language (e.g., syntactic class, abstract vs. concrete words). Thus, the arbitrariness, systematicity and iconicity of a given word is a continuous property rather than a categorical one. Chapter 4 explores whether words that are relatively more iconic are easier to process. Chapter 5 will consider what kinds of words tend to be higher in iconicity.

The Studies to Follow

In the subsequent chapters I explore two broad questions related to sound symbolism and iconicity. In Chapters 2 and 3 I approach the question of what creates sound symbolic associations. Then, in chapters 4 and 5, I turn my attention towards iconicity, and the effects of this phenomenon on language.

Chapter 2

Sound symbolic associations have been demonstrated for a variety of dimensions, and in a variety of populations. However, there is no consensus as to what causes these associations. In

Chapter 2 I begin to explore the origins of these associations by synthesizing the existing literature into five proposals for mechanisms underlying sound symbolic associations. In this chapter I will also provide a more detailed discussion of the topics presented in brief in this introduction.

Chapter 3

I then move to a series of studies that explore a novel sound symbolic association between phonemes and personality factors from the HEXACO model of personality (Lee &

Ashton, 2004). One goal of this work was to explore sound symbolism beyond the perceptual dimensions in which it is typically studied, and to explore whether associations could exist between phonemes and more abstract dimensions as well. While worthwhile in its own right, this question also contributes to an understanding of the mechanisms that could underlie sound

10 symbolism. By exploring the range of dimensions involved in sound symbolism, I develop a better understanding of the range of phenomena that mechanisms of sound symbolism must be able to explain.

Across three studies I show that certain kinds of phonemes (i.e., sonorants and voiceless stops) are associated with distinct factors from the HEXACO. In seeking to rule out some potential mechanisms for the effect I demonstrate these associations do not derive from real life

(e.g., individuals with sonorant names do not tend to be higher in the relevant personality factors) and are not mediated by likability. I consider how the mechanisms outlined in Chapter 2 might apply to this novel form of sound symbolism. This study also begins to move towards an understanding of the effects of sound symbolism in real language, by using existing names as stimuli.

Chapter 4

Moving from phoneme-feature associations, I turn my attention to form-meaning relationships in words. In particular, I explore the behavioural consequences of iconicity in language. Previous work has demonstrated that, compared to arbitrary words, iconic words: invoke different regions of the brain (e.g., Kanero, Imai, Okuda & Matsuda, 2014), show different patterns of neural activity (e.g., Peeters, 2016), are learned earlier (e.g., Perry, Perlman,

& Lupyan, 2015) and are easier for adults to learn (e.g., Lockwood, Hagoort, & Dinemanse,

2016). In Chapter 4 I present studies suggesting that iconic words are also easier to process, perhaps due to the special links between form and meaning.

Chapter 5

In light of the benefits that iconic words enjoy, one might wonder why language isn’t more iconic. Previous work has shown that there are limits to the kinds of meanings that can be

11 conveyed iconically (e.g., Winter, Perlman, Perry, & Lupyan, 2017). In Chapter 5 I explore two such limiting factors: the uniqueness of a word’s meaning and the amount of associated sensory information. I discover that it is words with unique meanings and a great deal of associated sensory information are most likely to have iconic forms. This demonstrates that iconicity exerts an influence on the makeup of the lexicon.

Chapter 2: Five Mechanisms of Sound Symbolic Association

Sound Symbolism

In this review, we use the term sound symbolism to refer to an association between phonemes and particular perceptual and/or semantic elements (e.g., large size, rounded contours).4 These associations arise from some quality of the phonemes themselves (e.g., their acoustic and/or articulatory features), and not because of the words in which they appear. Thus we exclude associations deriving from patterns of phoneme use in language (i.e., conventional sound symbolism; Hinton, Nichols, & Ohala, 1994) from our definition. We also exclude associations deriving from direct imitations of sound (i.e., imitative sound symbolism; Hinton et al., 1994).5 We exclude these associations because, like Hinton et al., we think they are distinct categories, and that they do not necessarily share underlying mechanisms with the phenomenon we seek to explain. As illustrated in Table 1, the definition of sound symbolism that we offer here is similar to a number of other definitions for the phenomenon that can be found in the literature.

4 While the term sound symbolism is used here at the phoneme level (i.e., involving relationships between individual phonemes and semantic elements), it has also been used at the word level (e.g., Johansson & Zlatev, 2013; Nielsen & Rendall, 2011; Tanz, 1971; Westbury, 2005). These two uses are not in opposition; sound symbolic words are those whose component phonemes have a sound symbolic relationship with their meanings. 5 Note that Hinton et al. (1994) used the terms conventional and imitative sound symbolism to refer to sound symbolism at the word level.

Table 1

Sample definitions of phonetic sound symbolism in the literature.

Sound Symbolism Definitions

“Sound symbolism is the process by which speakers link phonetic features with meanings non- arbitrarily” (D’Onofrio, 2013, p. 1).

“Synesthetic sound symbolism is the process whereby certain vowels, consonants, and suprasegmentals are chosen to consistently represent visual, tactile, or proprioceptive properties of objects, such as size or shape” (Hinton, Nichols, & Ohala, 1994, p. 4).

“Phonetic symbolism…proposes that an arbitrary linguistic sound itself carries symbolic weight, in that it evokes a sense of relatedness to other entities, such as color, touch, or emotion” (Hirata et al., 2011, p. 929).

“The idea of phonetic symbolism implies that sounds carry intrinsic symbolic connotations”

(Koriat & Levy, 1977, p. 93).

“The term sound symbolism is used when a sound unit such as a phoneme, syllable, feature, or tone is said to go beyond its linguistic function as a contrastive, non-meaning-bearing unit, to directly express some kind of meaning” (Nuckolls, 1999, p. 228).

“Sound symbolism refers to cases in which particular images are associated with certain sounds” (Shinohara & Kawahara, 2010, p. 1).

The term association is somewhat difficult to characterize in this context; broadly it refers to the sense that the phonemes in question seem related to, or to naturally go along with, stimuli possessing the associated elements or features (e.g., objects of a certain size or shape).

Sound symbolic associations emerge behaviourally in reports that nonwords containing certain

14 phonemes are especially good labels for particular targets (e.g., Maurer, Pathman, & Mondloch,

2006; Nielsen & Rendall, 2011). They may also emerge on implicit tasks, such that congruent phoneme-stimuli pairings are responded to differently than incongruent pairings (e.g., Hung,

Styles, & Hsieh, 2017; Ohtake & Haryu, 2013; Westbury, 2005).

These sound symbolic associations have important implications for our understanding of language. While the arbitrariness of language has long been considered one of its defining features (e.g., Hockett, 1963), sound symbolism allows one way for non-arbitrariness to play a role. It does this through congruencies between the sound symbolic associations of a word’s phonemes and the word’s meaning. An example of this could be when a word denoting something small contains phonemes that are sound symbolically associated with smallness (i.e., indirect iconicity, discussed later). These congruencies can have effects on language learning

(e.g., Asano et al., 2015; Imai, Kita, Nagumo, & Okada, 2008; Perry, Perlman, & Lupyan, 2015; for a review see Imai & Kita, 2014) and processing (e.g., Kanero, Imai, Okuda, Okada, &

Matsuda, 2014; Lockwood & Tuomainen, 2015; Sučević, Savić, Popović, Styles, & Ković,

2015). Moreover, sound symbolic associations have also been shown to impact cognition more broadly, including effects on action (Parise & Pavani, 2011; Rabaglia, Maglio, Krehm, Seok, &

Trope, 2016; Vainio, Schulman, Tiippana, & Vainio, 2013; Vainio, Tiainen, Tiippana, Rantala,

& Vainio, 2016), memory (Bankieris & Simner, 2015; Lockwood, Hargoort, & Dingemanse,

2016; Nygaard, Cook, & Namy, 2009; Preziosi & Coane, 2017) and categorization (Ković,

Plunkett, & Westermann, 2010; Lupyan & Casasanto, 2015; for a recent review of sound symbolism effects, see Lockwood & Dingemanse, 2015).

Interest in sound symbolism within psychology is on the rise. Ramachandran and

Hubbard’s (2001) article, which rekindled interest in the phenomenon6, was one of only 28 published on sound symbolism and/or the closely related topic of iconicity (discussed later) in that year. For comparison, a total of 193 articles were published on sound symbolism and/or iconicity in 2016 (see Figure 1). However, despite growing interest in the phenomenon, one topic that has largely been neglected is the mechanism underlying these associations. That is, mechanisms to explain why certain phonemes come to be associated with particular perceptual and/or semantic features. While there are a number of proposals, there is a scarcity of experimental work focused on adjudicating between them. One potential reason for this is that the mechanisms have yet to be thoroughly described and evaluated in a single work (though see

Deroy & Auvray, 2013; Fischer-Jørgensen, 1978; French, 1977; Johansson & Zlatev, 2013;

Masuda, 2007; Nuckolls, 1999; Shinohara & Kawahara, 2010); that is the aim of the present article. We begin by describing two well-known instances of sound symbolism to serve as reference points. Then, as an illustration of this topic’s importance, the role of sound symbolism in language is reviewed. Next we review the features of phonemes that may be involved in associations, and then explore the proposed mechanisms by which these features come to be associated with particular kinds of stimuli. Finally, we identify the outstanding issues that need to be addressed on this topic, and suggest potential next steps for the field.

6 The topic itself dates back at least to the fifth century BC, when Plato’s Cratylus takes place. This dialogue discusses the origin of words and contrasts a conventionalist perspective (i.e., that convention alone dictates the forms of words) with a naturalist perspective (i.e., that forms are naturally well suited for particular referents). These were popular topics of debate at the time (Sedley, 2013). It also includes interesting sound symbolic proposals, for instance /n/ being an internal sound, fit for meanings such as within or inside.

Percentage of Psychology Publications on Sound Symbolism and/or Iconicity

0.12

0.1

0.08

0.06

0.04 Percentage of Total Psychology Publications Psychology Total of Percentage 0.02

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Year of Publication

Figure 1. The percentage of psychological publications per year that included the term “sound symbolism” and/or “iconicity”, according to PsycINFO.

Size and Shape Symbolism

The two most well known sound symbolic effects are typically traced to a pair of works from 1929 (though there are relevant earlier observations; e.g., Gabelentz, 1891; Jesperson,

1922; for a review see Jakobson & Waugh, 1979). One of these is the Mil/Mal effect (Sapir,

1929), referring to an association between high and front vowels (see Table 2), and small objects; and low and back vowels, and large objects (Newman, 1933; Sapir, 1929). That is, when individuals are asked to pair nonwords such as mil and mal with a small and a large shape, most will pair mil with the small shape and mal with the large shape. Beyond a number of such explicit demonstrations (e.g., Thompson & Estes, 2010), the effect has also been shown to emerge implicitly. Participants are faster to respond on an implicit association task (IAT) if

17 mil/small shapes and mal/large shapes share response buttons compared to when the pairing is reversed (Parise & Spence, 2012). In addition, participants are faster to classify a shape’s size if a sound symbolically congruent (vs. incongruent) vowel is simultaneously presented auditorily

(Ohtake & Haryu, 2013). The effect has been demonstrated across speakers of different languages (e.g., Shinohara & Kawahara, 2010) and at different points in the lifespan (e.g., in the looking times of four-month-old infants; Peña, Mahler, & Nespor, 2011).

Table 2

Definitions of linguistic terms used throughout the paper (derived from Ladefoged & Johnson,

2010; Reetz & Jongman, 2009).

Phoneme Term Examples

Affricate consonants involve a combination of stops and /tʃ/ as in chat, /dʒ/ as in jack fricatives.

Alveolar consonants involve the tip of the tongue /t/ as in tab, /d/ as in dab contacting the alveolar ridge.

Approximant consonants involve a minor constriction in /l/ as in lack, /w/ as in whack airflow that does not cause turbulence.

Back vowels are those articulated with the highest point of /u/ as in who’d, /ɑ/ as in hawed the tongue relatively close to the back of the mouth.

Bilabial consonants involve the lips coming together in /m/ as in mat, /b/ as in bat their articulation.

Fricative consonants involve a major constriction in /f/ as in fat, /v/ as in vat airflow that does cause turbulence.

Front vowels are those articulated with the highest point of /i/ as in heed, /æ/ as in had the tongue relatively close to the front of the mouth.

High vowels are those articulated with the tongue relatively /i/ as in heed, /u/ as in who’d close to the roof of the mouth.

Low vowels are those articulated with the tongue relatively /æ/ as in had, /ɑ/ as in hawed far from the roof of the mouth.

Nasal consonants involve airflow proceeding through the /m/ as in mat, /n/ as in gnat nose.

Obstruent consonants involve a stoppage of, or turbulence in, the airflow; this includes stops, fricatives and affricates.

Rounded vowels are those articulated with rounded lips. /u/ as in who’d, /oʊ/ as in hoed

Sonorant consonants involve no stoppage of, or turbulence in, the airflow; this includes nasals and approximants.

Stop consonants involve a stoppage of airflow. /p/ as in pat, /b/ as in bat

Unrounded vowels are those articulated without rounded /i/ as in heed, /æ/ as in had lips.

Velar consonants involve the back of the tongue contacting /k/ as in cap, /ɡ/ as in gap the soft palette.

Voiced consonants involve the vocal folds being brought /b/ as in bam, /d/ as in dam close enough together to vibrate.

Voiceless consonants involve the vocal folds not being /p/ as in pat, /t/ as in tat brought close enough together to vibrate

Another well-studied sound symbolic association is the Maluma/Takete effect (Köhler,

1929), referring to an association between certain phonemes and either round or sharp shapes.

More recently, this has often been called the Bouba/Kiki effect, referring to the stimuli used by

Ramachandran and Hubbard (2001) in their demonstration of the effect. In general, voiceless stop consonants (i.e., /p/, /t/ and /k/)7 and unrounded front vowels (e.g., /i/ as in heed) seem to be

7 Though note that the shape associations of the voiceless bilabial stop /p/ have been somewhat equivocal (see D’Onofrio, 2013; Fort et al., 2015).

20 associated with sharp shapes; while sonorant consonants (e.g., /l/, /m/ and /n/), the voiced bilabial stop consonant /b/, and rounded back vowels (e.g., /u/ as in who’d), are associated with round shapes (D’Onofrio, 2013; Nielsen & Rendall, 2011, Ozturk, Krehm, & Vouloumanos,

2013; cf. Fort, Martin, & Peperkamp, 2015). As with the Mil/Mal effect, the Maluma/Takete effect has been repeatedly demonstrated using explicit matching tasks (e.g., Maurer et al., 2006;

Nielsen & Rendall, 2011; Sidhu & Pexman, 2016). It also emerges on implicit tasks such as the

IAT (Parise & Spence, 2012), and on lexical decision tasks, such that nonwords are responded to faster when presented inside of congruent (vs. incongruent) shape frames (e.g., a sharp nonword inside of a jagged vs. curvy frame; Westbury, 2005; cf. Sučević et al., 2015). It has been demonstrated in speakers of a number of different languages (e.g., Bremner et al., 2013; Davis,

1961; cf. Rogers &, 1975) and in the looking times of four-month-old infants (Ozturk, Krehm, &

Vouloumanos, 2013; cf. Fort, Weiß, Martin, & Peperkamp, 2015; Pejovic & Molnar, 2016).

Arbitrariness and Non-Arbitrariness

Sound symbolism is relevant to our understanding of the fundamental nature of spoken language, in particular, to the relationship between the form of a word (i.e., its articulation, phonology and/or orthography) and its meaning. One possibility is that this relationship is arbitrary, with no special connection between form and meaning (e.g., Hockett, 1963).8 Hockett

(1963) described this lack of special connection as the absence of a “physical or geometrical resemblance between [form and meaning]” (p. 8). However this seems to only contrast arbitrariness with iconicity (see below). A more general way of characterizing this lack of a

8 Readers familiar with the sound symbolism literature will no doubt notice the absence of reference to Ferdinand de Saussure’s Course in Linguistics (1916), which famously stated that “the bond between the signifier and the signified is arbitrary” (p. 67). As reviewed in Hutton (1989), Saussure may have intended to use the term arbitrary to describe the relationship between the abstract, mentalistic entities of the signifier and signified, as opposed to the form of a word and its referent in the world. It is this latter sort of arbitrariness that is relevant to sound symbolism. See Joseph (2015) for further discussion of this and Saussure’s later work, which explored iconicity as a factor in language change.

21 special connection is that aspects of a word’s form cannot be used as cues to its meaning

(Dingemanse et al., 2015). As an illustration, it would be difficult to derive the meaning of the word fun from aspects of its form.9 An important related concept is conventionality, the notion that words only mean what they do because a group of language users have agreed upon a definition.

It is also possible for the relationship between form and meaning to be non-arbitrary, either through systematicity or iconicity (Dingemanse et al., 2015). Systematicity refers to broad statistical relationships among groups of words belonging to the same semantic or syntactic categories. For instance, Farmer, Christiansen, and Monaghan (2006) showed that English nouns tend to be more phonologically similar to other nouns than to verbs (and vice versa for verbs).

Similarly, Reilly and Kean (2007) demonstrated that there are general differences in the forms of concrete and abstract English nouns. Importantly, systematicity does not involve relationships between words’ forms and their specific meanings, but broad relationships between groups of words and linguistic categories (Dingemanse et al., 2015). For instance, the nouns member, prison and student are systematic in that they have a stress on their initial syllable (as do most disyllabic nouns; Sereno, 1986). This is a non-arbitrary property in that it is possible to derive grammatical category from word form. However, initial syllable stress is not related to these words’ specific meanings in any particular way. While systematicity tends to occur on a large scale within a language, specific patterns of systematicity vary from language to language

(Dingemanse et al., 2015).

9 Surprisingly, finding a word to exemplify arbitrariness was quite difficult. This is illustrative of the point to follow, that most words contain a combination of arbitrary, systematic, and iconic elements. We chose fun because of its low iconicity rating (Winter, Perry, Perlman, & Lupyan, 2017) and derived systematicity value (Monaghan, Shillcock, Christiansen, & Kirby, 2014). Its length is also atypical of abstract nouns, which tend to be longer than concrete ones (Reilly & Kean, 2007), though this raises the interesting question of whether anti-systematic words are arbitrary.

The other way that language can be non-arbitrary is through iconicity: a resemblance between form and meaning.10 Instead of a holistic resemblance, this often emerges as a structural, resemblance-based mapping between aspects of a word’s form and aspects of its meaning (Emmorey, 2014; Meir et al., 2013; Taub, 2001). For instance, consider the example used by Emmorey (2014), of the hand sign for bird in American Sign Language, in Figure 2.

Notice that specific features of the form map on to specific features of the meaning (e.g., the presence of a protrusion at the mouth, the ability of that protrusion to open vertically). Because only certain aspects of meaning are included in the mapping, there are elements of the concept bird that are not represented (e.g., its wings). An example of this in spoken language is onomatopoeia: words that sound like their referent. Take for instance the word ding, whose abrupt onset and fading offset map onto these features in the sound of a bell (Taub, 2001; cf.

Assaneo, Nichols, & Trevisan, 2011). The preceding examples could be considered instances of direct iconicity, in which form maps directly onto meaning via resemblance (Masuda, 2007).

This mapping is of course constrained by the form’s modality; spoken language affords direct mapping onto meanings related to (or involving) sound, while signed languages are able to directly map onto spatial and kinesthetic meanings (Meir et al., 2013; Perniss et al., 2010).

10 This discussion focuses on phonological iconicity, however it is also possible to have iconicity at the level of morphemes (e.g., the addition of a plural suffix making a word larger; Jakobson, 1965), syntax (e.g., word order resembling temporal order; Perniss, Thompson, & Vigliocco, 2010) and prosody (e.g., the tendency to use a faster rate of speech when discussing faster movements; Shintel, Nusbaum, & Okrent, 2006).

Figure 2. The sign for bird in American Sign Language. Notice that specific aspects of the word’s form map onto specific aspects of its meaning. For instance, the presence of a protrusion at the mouth, and the ability of that protrusion to open vertically. From “ASLU” by W. Vicars,

2015 (http://lifeprint.com/index.htm). Copyright [1997] by LifePrint Institute. Reprinted with permission.

It is also possible for language to display indirect iconicity in which it is the forms’ associations that map onto meaning (Masuda, 2007). This was put elegantly by Von Humboldt

(1836) as cases in which sounds “produce for the ear an impression similar to that of the object upon the soul” (p. 73). In indirect iconicity it is the impression of the sound that maps onto meaning as opposed to the sound itself.11 Consider for instance the word teeny (/tini/). Because its meaning is not related to sound, its phonemes cannot map onto meaning directly. However, as mentioned, the high-front vowel phoneme /i/ is sound symbolically associated with smallness.

Thus, this phoneme maps onto smallness indirectly, by way of its sound symbolic association,

11 This kind of iconicity is by its very nature subjective, dependent on the associations a person makes (for a discussion see Hutton, 1989; Joseph, 2015). Nevertheless when an association is salient enough that it is apparent to a large group of language users, it merits consideration as a genuine phenomenon.

24 allowing teeny to be indirectly iconic. This is the relevance of sound symbolism to language: it provides one mechanism by which words can be non-arbitrarily associated with their meanings.

The preceding examples of iconicity would be considered instances of imagic iconicity: a relationship between a single form and meaning (Peirce, 1974). However, some have proposed that sound symbolism plays a role in diagrammatic iconicity: cases in which the relationship between two forms resembles the relationship between their two meanings. Imagic and diagrammatic iconicity are sometimes referred to as absolute and relative iconicity, respectively

(e.g., Dingemanse et al., 2015). Diagrammatic iconicity is often seen in ideophones, a class of words that depict various sensory meanings (beyond sounds) through iconicity (see Dingemanse,

2012a). For instance, the Japanese ideophones goro and koro mean a heavy and a light object rolling, respectively. Note that goro begins with a voiced consonant while koro begins with a voiceless consonant; voiced (voiceless) consonants are associated with heaviness (lightness; Saji,

Akita, Imai, Kantartzis, & Kita, 2011). Thus the relationship between the sound symbolic properties of each word (i.e., one being sound symbolically heavier than the other) reflects the relationship between their meanings. At the moment it is unclear whether sound symbolism primarily contributes to indirect imagic iconicity, or requires the comparison inherent in diagrammatic iconicity (e.g., Gamkrelidze, 1974). In Figure A1 in Appendix A we propose a taxonomy of iconicity that is an attempt to synthesize the various distinctions that have been made in the literature.

There is a good deal of work demonstrating that iconicity is present in the lexicons of spoken languages.12 The clearest example of this is the widespread existence of ideophones.

Although they are rare in Indo-European languages, they are common in many others, including:

12 The presence of iconicity in signed languages is of course more obvious and less controversial (for a review see Perniss et al., 2010).

25 sub-Saharan African languages, Australia Aboriginal languages, Japanese, Korean, Southeast

Asian languages, South America indigenous languages and Balto-Finnic languages (Perniss et al., 2010). Additionally, speaking to the psychological reality of ideophones, studies have shown that there are both behavioural (e.g., Imai et al., 2008; Lockwood, Dingemanse, & Hagoort,

2016) and neural differences (e.g., Kanero et al., 2006; Lockwood et al., 2016; Lockwood &

Tuomainen, 2016) in the learning and processing of ideophones as compared to non-ideophonic words (or ideophones paired with incorrect meanings).

There is also evidence that iconicity plays a role in the lexicon beyond ideophones. For instance, Ultan (1978) found that among languages that use vowel ablauting to denote diminutive concepts, most do so with high-front vowels. This is an example of indirect iconicity, occurring via high-front vowels’ sound symbolic associations with smallness. In addition, Blasi,

Wichmann, Hammarström, Stadler and Christiansen (2016) compared the forms of 100 basic terms across 4298 languages, and found, in addition to other patterns, that words for the concept small tended to include the high-front vowel /i/. Cross-linguistic studies have also reported evidence of indirect iconicity in, among other things, proximity terms (e.g., Johansson & Zlatev,

2013; Tanz, 1971), singular vs. plural markers (Ultan, 1978) and animal names (Berlin, 1994).

The ability of individuals to guess the meanings of foreign antonyms at an above chance rate

(e.g., Bankieris & Simner, 2015; Brown, Black, & Horowitz, 1955; Klank, Huang, & Johnson,

1971) has been attributed to indirect iconicity.

Taking an even broader view of language, Perry et al. (2015) and Winter et al. (2017) conducted a large scale norming study in which 3,001 English words were rated on a scale with zero indicating arbitrariness, and five indicating iconicity. The average rating was significantly greater than zero, indicating that this sample of words was not entirely arbitrary. Moreover, the

26 iconicity of words in this sample is related to age of acquisition (Perry et al., 2015), frequency, sensory experience (Winter et al., 2017), and semantic neighbourhood density (Sidhu & Pexman,

2018b). Thus, instead of being a linguistic curiosity, iconicity appears to be a general property of language that behaves in a predictable manner, even in a less obviously iconic language such as

English.

Of course, the existence of systematicity and iconicity does not discount the premise that arbitrariness is a fundamental property of language. As put by Nuckolls (1999), “throughout the exhaustive dissections and criticisms of the principle of arbitrariness, there has never been a serious suggestion that it be totally abandoned” (p. 246). Instead, arbitrariness, systematicity and iconicity are seen as three coexisting aspects of language (Dingemanse et al., 2015). In fact, there is a growing appreciation that words do not fall wholly into the categories arbitrary and non- arbitrary, but rather that individual words can contain both arbitrary and non-arbitrary elements

(e.g., Dingemanse et al., 2015; Perniss et al., 2010; Waugh, 1992). For instance, consider the word hiccups. It is a noun with a stressed first syllable (a systematic property); it also imitates aspects of its meaning (an iconic property). However, without knowing its definition, one would not be able to fully grasp its meaning based solely on its form (an arbitrary property). It seems that each of these properties contribute to language in varying proportions; they each also provide unique benefits to language. That is, systematicity facilitates the learning of linguistic categories (e.g., Cassidy & Kelly, 1991; Fitneva, Christiansen, & Monaghan, 2009; Monaghan et al., 2011). Iconicity makes communication more direct and vivid (Lockwood & Dingemanse,

2015), and can facilitate language learning (e.g., Imai et al., 2008; for a review see Imai & Kita,

2014). Lastly, decoupling form and meaning (i.e., arbitrariness) allows language to denote

27 potentially limitless concepts (Lockwood & Dingemanse, 2015), and avoids confusion among similar meanings with similar forms (e.g., Gasser, 2004; Monaghan et al., 2011).

Phonetic Features Involved in Sound Symbolism

Before turning to a discussion of how sound symbolic associations between phonemes and particular stimuli arise, it is important to make clear that in the present review we conceptualize these associations as arising from associations between specific phonetic features13 and particular perceptual and/or semantic features. For instance, the association between high- front vowels and smallness (i.e., the Mil/Mal effect) is seen as arising from an association between some component acoustic or articulatory feature of high-front vowels, and smallness.

Phonemes are multidimensional bundles of acoustic and articulatory features, any or all of which may afford an association with particular stimuli (e.g., Tsur, 2006). Indeed, Jakobson and Waugh

(1979) opine that “[m]ost objections to the search for the inner significance of speech sounds arose because the latter were not dissected into their ultimate constituents” (p. 182). Thus, the first step is to delineate these various features of vowel and consonant phonemes that may be involved in associations.

Vowels are phonemes produced by changing the shape and size of the vocal tract, through moving the tongue’s position in the mouth and opening the jaw. This is done without obstructing the airflow–without having the articulators (e.g., tongue, lips) come together. The three main articulatory features that determine vowels’ identity are: height (proximity of the tongue body to the roof of the mouth), frontness (proximity of the highest point of the tongue to

13 We use the term features more broadly than it would be used in the context of a strict phonological analysis (e.g., Jakobson, Fant, & Halle, 1951). Our discussion of features is also less exhaustive than would be found in such a context.

28 the front of the mouth) and lip rounding (see Figure 3).14 Vowels are described acoustically in terms of their formants: bands of high acoustic energy at particular frequencies. Tongue and jaw position serve to change the configuration of the vocal tract and affect which frequencies will resonate most strongly. The lowest of these formants (i.e., fundamental frequency) corresponds with the pitch of a vowel; it tends to be higher for high vowels, potentially because the tongue’s height “pulls on the larynx and thus increases the tension of the vocal cords” (Ohala & Eukel,

1987, p. 207), increasing pitch. The next three formants determine the identity of a vowel. The frequency of the first formant (F1) is negatively correlated with height; the frequency of the second formant (F2) is positively correlated with frontness. These relationships are due to changes in the volume of resonating cavities in the vocal tract when the articulators are in different positions. In addition, lip rounding lowers the frequency of all formants above the first

(in particular, the third). The distance between these formants (i.e., formant dispersion) is also important. For instance, front vowels are characterized by larger formant dispersion (between F1 and F2) than back vowels.

14 Another distinction is between tense (e.g., /i/ and /u/) and lax (e.g., /ɪ/ and /ʊ/) vowels. As noted by Ladefoged and Johnson (2010), this distinction is not simply based on muscular tension in their articulation; instead, the language- specific contexts in which they can appear differ. For instance, in English content words, tense vowels can appear in open syllables (e.g., bee, boo), while lax vowels cannot. While we have eschewed discussion of this in the main text in favour of dimensions that are more often discussed in the sound symbolism literature, there is some evidence of tenseness being involved in sound symbolism (e.g., Greenberg & Jenkins, 1966). Moreover, the tense/lax distinction is related to vowel length, with tense vowels tending to be longer than lax vowels (Ladefoged & Johnson, 2010); some studies have indeed implicated vowel length in sound symbolism (e.g., Newman, 1933).

Figure 3. An illustration of vowel space. The x-axis corresponds to the front-back dimension; the y-axis corresponds to the high-low dimension.

In articulating consonants, the airstream is obstructed in some way; consonants are defined based on the manner of this obstruction and the place where it occurs (Ladefoged &

Johnson, 2010). Broadly speaking, consonants’ manner of articulation can be divided into obstruents (produced with a severe obstruction of airflow) and sonorants (produced without complete stoppage of, or turbulence in, the airflow; Reetz & Jongman, 2009). Obstruents include stops (in which airflow is entirely blocked and then released in a burst), fricatives (in which airflow is made turbulent by bringing two articulators together) and affricates (a combination of the two). Sonorants include nasals (in which airflow proceeds through the nasal cavity) and approximants (in which airflow is affected by bringing two articulators together, though not enough to create turbulence). In the case of obstruent consonants, they can also be distinguished by whether the vocal folds are brought close together enough to vibrate (i.e., voiced consonants) or not (i.e., voiceless consonants); sonorant consonants are typically voiced. Place of articulation

30 refers to the location at which the airflow is affected, especially relevant categories include: bilabials (in which the lips are brought together), alveolars (in which the tongue tip is brought to the alveolar ridge) and velars (in which the back of the tongue is brought to the soft palate).

As with vowels, each of these articulatory features of consonants have acoustic consequences. Stops involve a period of silence (potentially with voicing) followed by a burst of sound as they are released (potentially with aspiration). Fricatives cause turbulent noise in higher frequencies; nasals involve formants similar to vowels, though much fainter, while approximants have stronger formant structures.

Consonants and vowels also affect one another through co-articulation. That is, very few words involve a single phoneme. The gestures involved in producing sequences of phonemes are quick and result in adjacent sounds influencing the articulation of one another. For instance, vowels can affect consonants’ formant transitions (an acoustic cue to the place of articulation). In addition, a vowel’s pitch can be affected by the consonant that precedes it (e.g., higher when preceded by a voiceless obstruent; Kingston & Diehl, 1994).

Mechanisms for Associations Between Phonetic and Semantic Features

Next we turn to the main topic of this review: how these phonetic features come to be associated with particular kinds of stimuli. This discussion will draw heavily from the literature on crossmodal correspondences which, broadly speaking, can be defined as “the mapping that observers expect to exist between two or more features or dimensions from different sensory modalities (such as lightness and loudness), that induce congruency effects in performance and often, but not always, also a phenomenological experience of similarity between such features”

(Parise & Spence, 2013, p. 792; also reviewed in Parise, 2016; Spence, 2011). For instance, individuals more readily associate bright objects with high-pitched sounds than low-pitched

31 sounds (Marks, 1974), and are faster to respond to objects if their brightness is congruent with a simultaneously presented tone (Marks, 1987). Our grouping of proposed explanations owes much to Spence’s (2011) grouping of proposed mechanisms for such crossmodal correspondences.

As noted by Parise (2016), the term crossmodal correspondence has been used to refer to associations between simple unidimensional stimuli, consisting of a single basic feature (e.g., pitch of pure tones, brightness of light patches), as well as associations between more perceptually complex, multidimensional stimuli, composed of multiple features from different modalities (e.g., linguistic stimuli, which contain multiple acoustic and articulatory features). If one considers crossmodal correspondences to encompass all associations between stimuli in different modalities, then sound symbolic associations would certainly fall into this category (as in Parise & Spence, 2012; Spence, 2011). However, associations involving either simple or complex stimuli could potentially be distinct phenomena (see Parise, 2016). Thus, in the following review, we use the term crossmodal correspondence only to refer to associations between basic perceptual dimensions (e.g., brightness and pitch), which make up the majority of the term’s usage (Parise, 2016). This draws a distinction between sound symbolic associations and crossmodal correspondences. Because phonemes are multidimensional stimuli, sound symbolism would be considered a distinct, though related, phenomenon from crossmodal correspondences. Thus, while mechanisms invoked to explain crossmodal correspondences can be informative, we must be cautious when extending them to sound symbolic associations.

In the following sections we group proposed explanations for sound symbolic associations into themes; note that although we think this grouping is helpful there may be instances in which a given explanation could fit under multiple themes. Additionally, while we

32 have included the themes that we feel best represent the existing literature, we acknowledge the possibility that other mechanisms may exist.

Mechanism 1: Statistical Co-Occurrence

One mechanism proposed to explain associations between sensory dimensions is the reliability with which they co-occur in the environment (see Spence, 2011). That is, experiencing particular stimuli co-occurring in the world may lead to an internalization of these probabilities.

This typically involves stimuli from a particular end of Dimension A tending to co-occur with stimuli from a particular end of Dimension B. One way of framing this is through the modification of Bayesian coupling priors, and the belief one has about the joint distribution of two sensory dimensions based on prior experience (Ernst, 2007).

Statistical co-occurrence has been proposed to explain the crossmodal correspondence between high (low) pitch and small (large) size (e.g., Gallace & Spence, 2006), due to the fact that smaller (larger) things tend to resonate at higher (lower) frequencies (see Spence, 2011).

Another example is the association between high (low) auditory volume and large (small) size

(e.g., Smith & Sera, 1992), which may arise from the fact that larger entities tend to emit louder sounds (see Spence, 2011). The plausibility of this mechanism has been demonstrated experimentally, by artificially creating co-occurrences between stimuli. Ernst (2007) presented participants with stimuli that systematically co-varied in stiffness and brightness (e.g., for some participants, stiff objects were always bright). After several hours of exposure participants demonstrated a crossmodal correspondence between these previously unrelated dimensions.

Further evidence comes from a neuroimaging study that showed that after presenting participants with co-occurring audio-visual stimuli, the presentation of stimuli in one modality was associated with activity in both auditory and visual regions (Zangenehpour & Zatorre, 2010).

This mechanism has been used to explain several sound symbolic associations. In these proposals, some component feature of the phonemes is claimed to co-occur with related stimuli in the environment. The most obvious application is to the Mil/Mal effect (see Spence, 2011). As mentioned, small (large) things tend to resonate at a high (low) frequency. Thus, front vowels may be associated with smaller objects because of front vowels’ higher frequency F2. Similarly, the association between high vowels and smaller objects may be due to high vowels’ higher pitch

(Ohala & Eukel, 1987).15 A similar explanation has also been proposed for the association between front (back) vowels and short (long) distances (Johansson & Zlatev, 2013; Rabaglia et al., 2016; Tanz, 1977). Tanz (1977) proposed that this might be due to the fact that lower frequencies are able to travel longer distances, and are therefore more likely to be heard from far away. Thus we often experience more distant entities co-occurring with lower frequency sounds, leading to an association between back vowels (which have a lower F2) and long distance.

The mechanism of statistical co-occurrence has also been applied to internally experienced co-occurrences. For instance, Rummer et al. (2014; also see Zajonc, Murphy, &

Inglehart, 1989) proposed that some phonemes might develop associations with particular emotions due to an overlap between the muscles used for articulation and those used for emotional expression. Previous research suggested that simply adopting the facial posture of an emotion can facilitate experience of that emotion (i.e., the facial feedback hypothesis, Strack,

Martin & Stepper, 1988). Rummer et al. (2014) noted that articulating an /i/ involves contracting the zygomaticus major muscle which is also involved in smiling; conversely, articulating an /o/

15 These explanations contain an element of indexicality, one of Peirce’s (1974) three sign elements, along with iconicity and symbolism (i.e., wholly arbitrary relationships). Indexes are defined by a relationship of contiguity between sign and object (e.g., smoke is an index of fire). Thus we might think of high frequencies being indexical of small size. This poses an interesting question regarding whether sound symbolism should only be discussed in relation to iconicity.

(as in the German hohe) involves contracting the orbicularis oris muscle, which blocks smiling.

They proposed that over time, the increased positive affect felt while articulating /i/ (due to facial feedback) will lead to that phoneme becoming associated with positive affect. Indeed, they showed that participants found cartoons funnier while articulating an /i/ as opposed to an /o/.

However, they did not directly examine facial feedback as a mechanism. In addition, the validity of the facial feedback hypothesis has recently been called into question by failures to replicate

Strack et al.’s original finding (Wagenmakers et al., 2016). Nevertheless, the notion that co- occurrences of phonemes and internal sensations can lead to sound symbolic associations is a possibility that invites further evaluation.

One final statistical co-occurrence account is worth mentioning, despite the fact that it is not presented as an account of sound symbolism. Gordon and Heath (1998) reviewed findings that several vowel shifts (systematic changes in how vowels are articulated in a population) seem to be moderated by gender, with females leading raising and fronting changes, and males leading lowering and backing changes. The term raising, for instance, refers to a given vowel being articulated with the tongue in a higher position than previously. They theorized that the different vocal tracts of women and men (contributing to women naturally having larger F2-F1 dispersion) might create an association between females and high-front vowel space (which has larger F2-F1 dispersion), and males and low-back vowel space. Females and males might then be drawn to gender stereotypical vowel space, leading to gender moderated vowel changes.16 Although the authors do not mention it, there is some evidence of a sound symbolic association between high- front vowels (low-back vowels) and femininity (masculinity; Greenberg & Jenkins, 1966; Tarte,

16 In addition to this explanation, the authors do also mention the possibility of females and males being drawn to different areas of vowel space because of the sound symbolic associations of those areas.

1982; Wu, Klink, & Guo, 2013; cf. Sidhu & Pexman, 2015). One might speculate that the natural co-occurrence between sex and formant dispersion contributes to this association.

There is a good deal of work that needs to be done to demonstrate that statistical co- occurrence is a viable mechanism for sound symbolism. The experimental evidence demonstrating that it can indeed create crossmodal correspondences (e.g., Ernst, 2007) makes it a promising mechanism. However, this evidence has been provided in the context of simple sensory dimensions; what remains to be seen is if such correspondences can then contribute to sound symbolic associations. That is, can a co-occurrence-based association between a component feature of a phoneme and certain stimuli, create a sound symbolic association for that phoneme as a whole? One way to examine this question would be to present participants with isolated phonetic components (e.g., high vs. low frequencies) co-occurring with perceptual features (e.g., rough vs. smooth textures). Experimenters could then examine if this co- occurrence led to a sound symbolic association between phonemes containing said phonetic components (e.g., phonemes with a high vs. low frequency F2) and targets containing said perceptual feature (e.g., rough vs. smooth textures). Another approach would be to interfere with existing associations by presenting stimuli that contradict them (e.g., large objects making high- pitched noises) and then examining the effect on sound symbolic associations.

An important feature of this mechanism is that it requires experience, and thus assumes that at least some sound symbolic associations are not innate (though, as will be discussed later, there are theories regarding evolved innate sensitivities to, and/or predispositions to acquire associations based on, certain statistical co-occurrences). As such, we might not expect associations that depend on statistical co-occurrences to be present from birth. Although Peña et al. (2011) found evidence for the Mil/Mal effect in four-month-old infants, it is possible that

36 even these very young infants had already begun to gather statistical information about the environment (see Kirkham, Slemner, & Johnson, 2002). Testing infants at an even younger age could allow us to investigate if less exposure to statistical co-occurrences results in a weaker sound symbolism effect (or the absence of an effect altogether). Of course, any differences between younger and older infants could simply be attributable to differences in cognitive development. Thus, another approach could be to test infants of the same age for associations based on co-occurrences that they are more or less likely to have experienced. For instance, young infants may have more experience of certain frequencies co-occurring with different sizes than with different distances; the effects of these differences in experience could be tested. Also, we would only expect associations of this kind to be universal if they are based on a universal co-occurrence. While natural co-occurrences reflecting physical laws (e.g., between pitch and size) may be relatively universal, it might be possible to find others that vary by location. For instance, some have speculated that advertising can create statistical co-occurrences that are relatively local, and that these potentially contribute to cultural variations in some crossmodal correspondences (e.g., Bremner et al., 2013). It could be informative to examine instances in which populations differ in culturally-based statistical co-occurrences, and to compare their demonstrated associations. As mentioned by Wan et al. (2014), one might also consider effects of geographical differences (e.g., in landscape or vegetation) on statistical co-occurrences.

Mechanism 2: Shared Properties

Another broad class of accounts includes proposals that phonemes and associated stimuli may share certain properties, despite being in different modalities. Again, these properties in phonemes would likely derive from one or more of their component features. Individuals may then form associations based on these shared properties. These explanations can be divided into

37 those involving low-level properties (i.e., perceptual) and those involving high-level properties

(i.e., conceptual, affective, or linguistic).

Low-level properties. Some perceptual features may be experienced in multiple modalities. For instance, one can experience size in both visual and tactile modalities. One way of explaining sound symbolic associations is to suggest that they involve an experience of the same perceptual feature in both phonemes and associated stimuli. For instance, Sapir (1929; see also Jesperson, 1922) theorized that participants might have associated high vowels with small shapes in part because for high vowels the oral cavity is smaller during articulation. Thus both phonemes and shapes had the property of smallness. Similarly, Johansson and Zlatev (2013) proposed this as one potential explanation for the association between high-front vowels and small distances. Many have also pointed out that the vowels associated with roundness (i.e., /u/ and /oʊ/ as in hoed) involve a rounded articulation (e.g., French, 1977; Ikegami & Zlatev, 2007; also suggested in Ramachandran & Hubbard, 2001). Note that these accounts involve some amount of abstraction, or other mechanism by which features can be united across modalities, and don’t necessarily imply that phonemes and stimuli possess identical perceptual features.

Nevertheless, they do imply a certain amount of imitation between phonemes and associated features.

Others have proposed similar, though less direct, accounts. For instance, Saji et al. (2011) theorized that the association between voiced (voiceless) consonants and slow (fast) actions has to do with the shared property of duration. That is, in voiced consonants the vocal cords vibrate prior to stop release, and thus for a longer time than in voiceless consonants. This longer duration might unite them with slow movements, which take a longer time to complete. Ramachandran and Hubbard (2001) also speculated that the Maluma/Takete effect might owe to an abruptness,

38 or “sharp inflection” (p. 19) in both voiceless stops and sharp shapes. Indeed, voiceless stops involve a complete absence followed by an abrupt burst of sound; similarly, the outlines of sharp shapes involve abrupt changes in direction.

One final proposal is that a phoneme may be associated with body parts highlighted in its articulation (this was originally suggested by Greenberg, 1978). This account stands out from those discussed elsewhere in this review in that associations purported to derive from it have not been demonstrated experimentally, but rather inferred from comparisons across languages. For instance, Urban (2011) found that across a sample of languages, words for nose and lips were more likely to contain nasals and labial stops, respectively, than a set of control words. In addition, Blasi et al. (2016) found that words for tongue tended to include the phoneme /l/ (for which the airstream proceeds around the sides of the tongue), while words for nose tended to include the nasal /n/. Importantly, the patterns documented by Blasi et al. did not seem to be a result of shared etymologies or areal dispersion; thus, the authors speculated that they could potentially have derived from sound symbolic associations (or a related phenomenon). If the association between phonemes and body parts that these findings seem to hint at exists, it would be much more direct and limited than other associations discussed in this review. Future behavioural studies might examine if, beyond these quasi-imitative relationships, phonemes are also associated with stimuli that are related to the relevant body part17 (e.g., nasals and objects with salient odours). Such associations could ostensibly derive from the shared property of a salient body part.

High-level properties. Others have proposed that the shared properties that produce sound symbolism are more conceptual in nature. For instance, Walker, Walker, and Francis

17 The phonestheme sn-, which appears in words related to the nose and mouth (e.g., snarl, sneeze, sniff; see the Language Patterns section), may be indicative of such an association (e.g., Waugh, 1993).

(2012) suggested that crossmodal correspondences might emerge due to shared connotative meaning (i.e., what the stimuli suggest, imply or evoke) among stimuli. Note that this is distinct from what the stimuli denote (i.e., what they directly represent). That is, a bright object denotes visual brightness, but this is distinct from a connotation of brightness, which can apply across modalities. For example, tastes and melodies can seem “bright”.

When we consider the fact that these suprasensory properties can be shared by stimuli across modalities, it becomes apparent that shared connotations might explain a wide variety of observed crossmodal correspondences. As an example, consider that high-pitched tones have the connotations of being brighter, sharper, and faster than low-pitched tones (Walker et al., 2012).

These connotations of high-pitched tones might explain the association between high pitches and small stimuli (which also share these connotations). Moreover, Walker and Walker (2012; see also Karowski, Odbert, & Osgood, 1942) proposed that there are a set of aligned connotations, such that a stimulus possessing one of them will also tend to possess the others. For instance, stimuli with the connotation of brightness will also tend to have connotations of sharpness, smallness and quickness (Walker et al., 2012).

This framework may extend to sound symbolic associations (see Walker, 2016). That is, some sound symbolic associations might arise due to phonemes and stimuli sharing connotations. The connotations of phonemes would derive from the connotations of their component features. For instance, high-front vowels, which are high in frequency, have the same connotations as high frequency pure tones (e.g., brighter, sharper, and faster). This might explain their association with small stimuli, which, as reviewed above, also share these connotations. In a test of this proposal, French (1977) hypothesized, and then investigated, a sound symbolic association between high-front vowels and coldness, based on a similarity in connotation

40 between coldness and smallness. Indeed, his participants reported that nonwords containing the vowel /i/ were the “coldest” while those containing /ɑ/ (as in hawed) were the “warmest”.

Similar explanations have also been applied to shape sound symbolism. Bozzi and Flores

D’Arcais (1967) asked participants to rate compatibility between nonwords and shapes, and also to rate both kinds of stimuli on semantic differential scales (i.e., Likert scales anchored by polar adjectives, used to measure connotations). They found that compatible nonwords and shapes tended to have similar connotations (e.g., sharp nonwords and shapes were both rated as being fast, tense and rough). Gallace, Boschin, and Spence (2011) made a similar proposal to explain their finding that round and sharp nonwords were differentially associated with certain tastes.

They found that these associations were predicted by similar ratings of nonwords and tastes on connotative dimensions such as tenseness or activity.

A limitation of this account is that it begs the question of how phoneme features come to be associated with their connotations. There are also several conceptual clarifications required.

For instance, in cases of several shared connotations, is one primary in creating the association?

In addition, there is a need to clarify the distinction between a given phoneme’s connotations and its sound symbolic associations. That is, when participants rate a given vowel as belonging to the

“small” end of a large/small semantic differential scale, does that describe a connotation, an associated perceptual (i.e., denotative) feature, or both? Should connotations themselves be considered instances of sound symbolism? The exact connotative dimensions involved also require further elaboration. Much of Walker’s work focuses a core set of connotations including: light/heavy, sharp/blunt, quick/slow, bright/dark, and small/large (e.g., Walker & Walker, 2012).

Others have focused on connotations that comprise the three factors of connotative meaning

41 discovered by Osgood, Suci, and Tannenbaum (1957), namely: evaluation (e.g., good/bad), potency (e.g., strong/weak) and activity (e.g., active/passive; e.g., Miron, 1961; Tarte, 1982).

It has also been proposed that some crossmodal correspondences arise via transitivity

(e.g., Deroy, Crisinel, & Spence, 2013). That is, if there exists an association between

Dimensions A-B, and B-C, this might create an association between Dimensions A-C. French

(1977) made a similar suggestion for sound symbolic associations. He theorized that phonemes are only directly related with a small number of stimulus dimensions, and that these mediate relationships with other stimulus dimensions. For instance, high-front vowels may be directly associated with smallness, which mediates a relationship between high-front vowels and other properties related to smallness (e.g., thinness, lightness, quickness). A clarification to make going forward is whether these mediated effects involve relationships among denotative (as in

Deroy et al., 2013) or connotative dimensions, or both (see Figure 4).

Figure 4. An illustration of the different ways in which denotative and/or connotative dimensions may result in mediated relationships. A) The sort of mediation discussed by Deroy et al. (2013) in which the perceptual dimension of smallness might mediate a relationship between high/front vowels and the perceptual dimension of quickness via transitivity. B) A relationship based off of

Walker & Walker (2012; which might be considered a mediated one) in which the connotation of smallness might mediate a relationship between high/front vowels and the perceptual dimension of quickness. C) An example of mediation involving both denotative and connotative dimensions, in which high/front vowels are associated with the perceptual dimension of quickness because of the phonemes’ association with the perceptual dimension of smallness, and

43 that dimension’s association with the perceptual dimension of quickness (via a shared connotation).

A related proposal is that stimuli may be associated by virtue of having the same impression on a person. That is, instead of being united through a shared conceptual property, stimuli may be associated because they have a similar effect on a person’s level of arousal or affect (Spence, 2011). Indeed there is some evidence of hedonic value (Velasco, Woods, Deroy,

& Spence, 2015) and associated mood (Cowles, 1935) underlying crossmodal correspondences.

This account has not yet been examined in the context of sound symbolism. However, as is discussed elsewhere in this review, there has been some work proposing a link between phonemes and particular affective states (e.g., Nielsen & Rendall, 2011, 2013; Rumner et al.,

2014).

Lastly, some have theorized that crossmodal correspondences arise when the two dimensions share the same labels (e.g., Martino & Marks, 1999). For instance, the correspondence between pitch and elevation may derive from the use of the labels high and low for both. Evidence for this has come from the fact that speakers of languages using different labels for pitch (e.g., high/low in Dutch; thin/thick in Farsi) show different crossmodal correspondences (e.g., height and pitch in Dutch speakers; height and thickness in Farsi speakers;

Dolscheid, Shayan, Majid, & Casasanto, 2013). Although this has not yet been proposed for sound symbolic associations, there are some relevant observations. For instance, front and back vowels are sometimes referred to as bright and dark vowels, respectively (e.g., Anderson, 1985).

This corresponds to the visual stimuli with which either group of phonemes is associated

(Newman, 1933). However, this is example only intended to serve as an illustration; at the moment the relevance of this account to sound symbolism is purely speculative. In addition, a

44 question related to this general explanation is one of directionality: do shared linguistic labels create associations, or vice versa, or both? Dolscheid et al. (2013) demonstrated that teaching

Dutch speakers to refer to pitch in terms of thickness led to effects that resembled those of Farsi speakers. Speaking to the converse, Marks (2013) theorized that crossmodal correspondences might contribute to the creation of linguistic metaphors, and the use of a term from one sensory modality to describe sensations in another (see also Shayan, Ozturk, & Sicoli, 2011).

An important step in testing theories based on shared properties will be demonstrating the involvement of the hypothesized shared properties (see Table 3 for a summary of properties).

With regards to conceptual properties, a potential starting point could be to examine ratings on semantic differential scales for phonemes and associated stimuli, to test whether they indeed share connotations. The next step could be to verify activation of the shared conceptual properties, potentially by examining if they become more accessible following sound symbolic matching. For affect-based associations, one could examine whether phonemes and associated stimuli elicit comparable changes in a person’s self-reported mood. Deroy et al. (2013) theorized that mediated relationships (involving denotative dimensions) should be weaker than direct ones.

This provides a potential way of detecting such relationships. In addition, several of these theories may depend on developmental milestones (e.g., the acquisition of language) and thus make different predictions for sound symbolic effects when individuals are tested before and after these milestones are reached. Lastly, associations based on shared properties would be expected to vary cross-culturally to the extent that stimuli differ in their associated properties across cultures.

Table 3

A summary of the shared properties that are proposed to be involved in sound symbolism.

Shared Property Example

Perceptual feature High-front vowels and small shapes sharing

the property smallness (Sapir, 1929)

Magnitude or intensitya Both high volume and brightness being high

in magnitude (see Spence, 2011)

Connotation Stop consonants and angular shapes having

the connotations of being fast and tense

(Bozzi & Flores D’Arcais, 1963)

Relationship with a mediating dimension High-front vowels being associated with

thinness via the mediating dimension of size

(French, 1977)

Affective quality and resulting impression Sweet taste and round shape being united

via their positive hedonic value (Velasco et

al., 2015)

Linguistic label Vowels referred to as bright or dark being

associated with high and low brightness,

respectively. a Magnitude and intensity are discussed in the section on Neural Factors.

An outstanding question that is important for these accounts is whether participants only recognize shared properties and form associations when asked to do so during a task. For

46 instance, when asked to rate the similarity between nonwords and tastes, participants might very well consider properties that the two have in common. However, this does not mean that such associations exist outside of that task context. One could address this issue by examining whether associations are detectable using implicit measures (e.g., priming) that do not force participants to consider the relationships between stimuli in an overt way. Indeed, Walker and

Walker (2012) demonstrated that a crossmodal correspondence based on connotation could affect responses on an implicit task.

Mechanism 3: Neural Factors

The third mechanism includes proposals that sound symbolic associations arise because of structural properties of the brain, or the ways in which information is processed in the brain.

To be clear, this is not to imply that other mechanisms do not rely on neural factors. The difference here is that the following theories propose neural factors to be the proximal causes of the associations.

A theory described in the crossmodal correspondence literature suggests that there may be a common neural coding mechanism for stimulus magnitude, regardless of modality. For instance, Stevens (1957) noted that increases in stimulus intensity result in higher neuronal firing rates. In a similar vein, Walsh (2003) proposed that a system in the inferior parietal cortex is responsible for coding magnitude, again across modalities. Thus, for stimulus dimensions that can be quantified in terms of more or less (e.g., more or less loud, more or less bright), this common neural coding mechanism may lead to an association between the “more” and the “less” ends of each dimension (see Spence, 2011). For instance, the correspondence between high (low) volume and bright (dim) objects (Marks, 1987) may have to do with the fact that they are both high (low) in magnitude (see Spence, 2011). So far this has not been extended to sound symbolic

47 associations. However, it may be a viable mechanism when involving phonetic features that can be characterized in terms of magnitude.18

Another relevant theory is based on a hypothesized relationship between the brain regions associated with grasping and with articulation. Some have proposed that the articulatory system originated from a neural system responsible for grasping food with the hands and opening the mouth to receive it, resulting in a link between articulation and grasping (see Gentilucci &

Corballis, 2006). Vainio et al. (2013) demonstrated that participants were faster to make a precision grip (i.e., thumb and forefinger) while articulating the phonemes /t/ or /i/, and faster to make a power grip (i.e., whole hand) while articulating the phonemes /k/ or /ɑ/. Note that the articulation of each set of phonemes reflects the performance of either kind of grip.19 Vainio et al. theorized that the Mil/Mal effect might emerge from these associations (see also Gentilucci &

Campione, 2011). For instance, seeing a small shape may elicit the simulation of a precision grip

(Tucker & Ellis, 2001), which would then also activate a representation of the phoneme /i/’s articulation. It should be noted, however, that a follow up study by this group found participants were no faster to articulate an /i/ (/ɑ/) in response to a small vs. large (large vs. small) target

(Vainio et al., 2016).20 Thus, there is still a need for more direct evidence of the proposed links.

An ideal way to examine these neural theories would be to use neuroimaging. For instance, it would be informative to test for activation in the hypothesized magnitude-coding

18 Of course, it is possible that this magnitude matching is not neurally based. For instance, Marks (1989) theorized that loud and bright stimuli might share a semantic code (i.e., be represented as intense). Thus, magnitude matching might be conceptualized as being based on the shared conceptual properties of high intensity or low intensity, as opposed to being fundamentally neural in origin. 19 The authors theorize that two separate, but potentially related, processes may be at work. The links between vowels and grips may be due to double grasp neurons: the mouth prepares to receive an object whose size is related to hand grip size. The links between consonants and grips may be due to a tendency to mirror hand movements with the speech musculature; for instance, note the similarity between the articulation of /t/ and a precision grip. 20 Interestingly though, participants were faster to articulate an /m/ or a /t/, in response to a round or a sharp shape, respectively.

48 region when processing phonemes and related stimuli. Likewise, testing for activation in motor regions associated with articulation, in response to graspable objects, could also provide insight into articulation/grasping as a neural mechanism. There is recent evidence for the converse relationship: increased activity in motor regions associated with performing a precision or power grip, while articulating /ti/ or /kɑ/, respectively (Komeilipoor, Tiainen, Tiippana, Vainio, &

Vainio, 2016). These mechanisms should be largely universal, and thus the neural accounts predict that sound symbolic associations should not be modulated by culture.

Mechanism 4: Species-General Associations

Some have explained sound symbolism as based on species-general, inherited associations. While other mechanisms may involve evolved processes, the following theories propose that the associations themselves (as opposed to the processes leading to those associations) are a result of evolution.

One of the most widely cited explanations for the Mil/Mal effect is Ohala’s (1994)

Frequency Code Theory. This is based on the observation that many non-human species use low- pitched vocalizations when attempting to appear threatening, and high-pitched vocalizations when attempting to appear submissive or non-threatening (Morton, 1977). Ohala proposes that these vocalizations appeal to, and are indicative of, an innate cross-species association between high (low) pitches and small (large) vocalizers (viz. the Frequency Code). Thus when an animal wants to appear threatening, they use a low-pitched vocalization in order to give off an impression of largeness. Ohala theorizes that humans’ association between frequency (e.g., in vowels’ fundamental frequency and F2) and size is due to this same Frequency Code. At a fundamental level this explanation is based on co-occurrence (i.e., between pitch and size); however, it is argued that sensitivity to this co-occurrence has become innate. As evidence for

49 this innateness, Ohala points to the fact that male voices lower at puberty: precisely when they will need to use aggressive displays (i.e., low-pitched vocalizations) to compete for a mate. He argues that such an elaborate anatomical evolution would only have been worthwhile if it appealed to an innate predisposition in listeners. Nevertheless, Ohala concedes that the

Frequency Code may require some post-natal experience of relevant environmental stimuli, to be fully developed. Thus, one might regard the Frequency Code Hypothesis as an innate predisposition to develop an association, rather than as an innate association per se.

It is important to note that while many studies have found a relationship between fundamental frequency and body size in several species (e.g., Bowling et al., 2017; Charlton &

Reby, 2016; Gingras, Boeckle, Herbst, & Fitch, 2003; Hauser, 1993; Wallschläger, 1980), others have not (e.g., Patel, Mulder, & Cardoso, 2010; Rendall, Kollias, Ney, & Lloyd, 2005; Sullivan,

1984). As noted by Bowling et al. (2017), a relevant factor seems to be the range in body sizes studied, with more equivocal effects when studying the relationships within a given category than across categories (e.g., within a species vs. across species; cf. Davies & Halliday, 1978;

Evans, Neave, & Wakelin, 2006). In response to these equivocal findings, Fitch (1997) presented results from research with rhesus macaques, demonstrating that formant dispersion may be a better indicator of body size than fundamental frequency. It is beyond the scope of this review to adjudicate between these two cues. However to the extent that formant dispersion is a more reliable cue to size than fundamental frequency, the Frequency Code Hypothesis may require reframing. It is relevant to note that the Mil/Mal effect can be characterized in terms of formant

50 dispersion, which is larger for front vowels than back vowels, and decreases from high-front vowels to low-front vowels.21

In a similar vein, Nielsen and Rendall (2011; 2013) note that many non-human species use harsh punctuated sounds in situations of hostility and high arousal; and smoother, more harmonic sounds in situations of positive affiliation and low arousal.22 Notably, the meanings of these calls do not need to be learned by conspecifics, suggesting an innate sensitivity to their meanings (Owren & Rendall, 2001). There is also evidence of this in humans: infants use harsh

(smooth) sounds in situations of distress (contentment); adults use harsh and punctuated voicing patterns in periods of high stress (Rendall, 2003). Nielsen and Rendall theorize that the evolved semantic-affective associates of these two types of sounds may extend to phonemes with similar acoustic properties: namely obstruents and sonorants. For instance, swear words (which can be considered threatening stimuli) contain a relatively large proportion of obstruents (Van Lancker

& Cummings, 1999). This could contribute to the Maluma/Takete effect, and to associations between stop phonemes (sonorant phonemes) and sharp (round) shapes. Such an account would depend on sharp shapes seeming more dangerous than round shapes, and indeed there is some speculation in this regard (Bar & Neta, 2006).

A potential limitation of the claims regarding evolved and/or innate traits is the challenge of generating testable hypotheses from these accounts. One approach would be to examine whether the relevant associations are present universally, and from a very young age. While there is evidence for sensitivity to the Mil/Mal effect (Peña et al., 2011) and the Maluma/Takete effect

21 Ohala (1994) mentions that vowel sound symbolism may depend on formant dispersion, citing Fischer-Jørgensen (1978), potentially suggesting that Ohala saw the Frequency Code Theory as compatible with a focus on formant dispersion. 22 These two accounts do seem to be related. As noted by Morton (1994), “Aggressive animals utter low-pitched often harsh sounds...appeasing animals use high-pitched, often tonal sounds” (p. 350).

(Ozturk et al., 2013) in four-month-old infants, it is notable that two other studies have failed to find evidence of infant sensitivity to the Maluma/Takete effect at that age (Fort et al., 2015;

Pejovic & Molnar, 2016). In addition, one might debate whether observing an effect at four months of age is sufficient to infer its innateness. Thus, the evidence for innateness is not overwhelming at present. At least one crossmodal correspondence has been demonstrated in infants between 20-30 days old (Lewkowicz & Turkewitz, 1980), and it would be informative for future studies to examine sensitivity to sound symbolism at a similar age. Another approach could be a comparative one, examining if non-humans demonstrate sound symbolism. Ludwig,

Adachi, and Matsuzawa (2011) reported a crossmodal correspondence between pitch and brightness in chimpanzees, suggesting that such an investigation might be worthwhile.

Mechanism 5: Language Patterns

One final group of theories proposes that sound symbolic associations emerge due to patterns in language. This is of course related to the first mechanism discussed (i.e., statistical co-occurrence); the important distinction is that, as opposed to observing co-occurrences in the environment, the theories to be discussed propose sound symbolic associations might derive from co-occurrences between phonological and semantic features in language. An example of this would be associations derived from phonesthemes: phoneme clusters that tend to occur in words with similar meanings (e.g., gl- in words relating to light, such as glint, glisten, glow; see

Bergen, 2004). After repeated exposure, individuals might come to associate /gl/ with brightness, for instance. Indeed there is evidence of individuals using their knowledge of phonesthemes when asked to generate novel words (e.g., using the onset gl- when asked to create a nonword related to brightness; Magnus, 2000). Bolinger (1950) even suggested that phonesthemes may

“attract” the meanings of semantically unrelated words that contain the relevant phoneme clusters, leading to semantic shifts towards the phonesthemic meaning.

Such proposals are typically presented as an explanation for a distinct subset of sound symbolism, and not as an explanation for sound symbolism as a whole (e.g., Hinton et al., 1994).

Indeed, our operational definition would consider associations arising in this manner to be a separate phenomenon altogether. Nevertheless, some have proposed that language patterns can explain all of sound symbolism (e.g., Taylor, 1963). This kind of proposal has, however, not been supported by large-scale corpus analyses. For instance, a study by Monaghan, Mattock, and

Walker (2012) did not find overwhelming evidence that certain phonemes tend to occur in meanings related to roundness or sharpness. This would seem to suggest that the Maluma/Takete effect cannot be explained by language patterns. We described some other instances of indirect iconicity in the lexicon earlier in this paper, but the fact that many of these instances emerge across large samples of languages leads to the conclusion that they are the result of sound symbolism as opposed to the cause of it (see also Blasi et al., 2016; Wichmann, Holman, &

Brown, 2010).

There is, however, a good deal of support for a weaker version of this claim, namely that language patterns modify and constrain sound symbolism. For instance, Imai and Kita (2014) proposed that young infants are sensitive to a wide variety of sound symbolic associations, but that associations not supported by the phonology of an infant’s language, or inventory of sound symbolic words, tend to fade away as the infant develops. This proposal is supported by evidence of a greater sensitivity to foreign sound symbolic words in children as compared to adults (Kantartzis, 2011). There is also evidence of a language’s phonology moderating sound symbolic associations for speakers of that language. A basic example of this is the finding that

53 individuals perceptually assimilate phonemes that don’t appear in their language into ones that do (e.g., Tyler, Best, Faber, & Levitt, 2014; see Best, 1995). Sapir (1929) theorized that this may have been the reason English speaking participants did not rate nonwords containing /e/ (as in the French été) as being as small as expected. Because this phoneme does not appear in English, participants may have projected onto it the qualities of the diphthong /eɪ/ (as in hay), which begins lower than /e/ for many speakers (Ladefoged & Johnson, 2010). Another example comes from a study by Saji et al. (2011), who found that high-back vowels were associated with slowness by Japanese speakers, but with quickness by English speakers. They theorized that this was due to the fact that this vowel is rounded in English but not in Japanese. Lastly, there is recent evidence that the distributional properties of phonemes in a given language can impact their tendency to show sound symbolic associations for speakers of that language (i.e., less frequent phonemes may be more likely to have sound symbolic associations; Westbury, Hollis,

Sidhu, & Pexman, 2017).

Contextual Factors

One final topic that deserves mention is the role of various contextual factors in sound symbolism. As in the weaker version of the language patterns theory outlined above, contextual factors likely moderate the expression of sound symbolic associations rather than create them.

For instance, some have theorized that forced choice tasks may lead participants to become aware of shared properties among stimuli that they would not have considered otherwise (e.g.,

Bentley & Varron, 1933; French, 1977). In addition, some authors have speculated that pairing sounds with congruent meanings in real language may serve to highlight potential associations

(e.g., Waugh, 1993). Dingemanse et al. (2016) point out that in some cases it is necessary to know the definition of a word in order to appreciate the sound symbolic association between its

54 phonemes and meaning. That is, would one appreciate the sound symbolism of goro without knowing that its definition related to heaviness? Tsur (2006) characterizes sound symbolic associations as “meaning potentials” (p. 917) that can be actualized by associating phonemes with meanings in language. As noted by Werner and Kaplan (1963), sounds demonstrate plurisignificance, in that they are able to be associated with multiple different dimensions. Tsur suggests that the semantic context in which words appear might highlight some potential associations over others. Lastly, prosody has been theorized to direct individuals towards particular sound symbolic associations (Dingemanse et al., 2016).

Another potential factor to consider is cultural variation in conceptualizations of the relationship between sound and meaning. Nuckolls (1999) reviewed case studies of a number of societies in which language sounds are seen as intimately related to the external world. For instance, the Navajo view air as a source of life, and manipulating that air in the service of creating linguistic sound as one way of making “contact with the ultimate source of life”

(Witherspoon, 1977, p. 61; Reichard, 1944, 1950). As another example, different states of water

(e.g., swirling, splashing) represent important landmarks for the Kaluli people of Papua New

Guinea (Feld, 1996). Their language contains a number of ideophones that depict these different states of water, representing a fascinating interplay of linguistic sound and geography. This interplay is exemplified in their poetry, which depicts waterways in both sound and structure.

Indeed, some have speculated that variations in ideophone usage may result from cultural variation in cognitive styles (e.g., Werner & Kaplan, 1963). One wonders if cultural factors may moderate the expression of sound symbolic associations.

Outstanding Issues and Future Directions

We have outlined five mechanisms that have been proposed to explain sound symbolic associations: the features of the phonemes co-occurring with stimuli in the environment, shared properties among phoneme features and stimuli, overlapping neural processes, associations created by evolution, and patterns extracted from language. There are a number of outstanding issues in the literature, and it is to these that we now turn our attention.

Phonetic Features

So far in this review we have been equivocal on whether sound symbolism involves acoustic or articulatory features. In fact, there is no need to attribute the phenomenon to one or the other; most theorists allow for both to potentially play a role (e.g., Newman, 1933; Nuckolls,

1999; Ramachandran & Hubbard, 2001; Sapir, 1929; Shinohara & Kawahara, 2010;

Westermann, 1927). This is commensurate with the notion of phonemes as bundles of acoustic and articulatory features, either/both of which can be associated with targets in sound symbolism

(e.g., Tsur, 2006).23 Indeed there is evidence of both playing a role. For instance, Tarte’s (1982) research comparing vowels to pure tones showed that vowels are associated with some stimuli in a way that would be expected if pairings were based on vowels’ component frequencies alone.

Eberhardt’s (1940) discovery of sound symbolism in profoundly deaf individuals suggested that articulatory features in isolation can contribute to sound symbolism (though admittedly in a specific population; cf. Johnson, Suzuki & Olds, 1964).

23 With this in mind, sound symbolism becomes something of a misnomer, as it seems to imply that acoustic features drive associations. Phonetic symbolism, a term that is sometimes used to refer to the same effect (see Spence, 2011; Table 1) might be more appropriate. However, we elected to use sound symbolism since it is the more common term (e.g., used exclusively 45 times in 2016, compared to 12 for phonetic symbolism, per PsycINFO).

While it seems reasonable to assume that both articulatory and acoustic features can play a role in sound symbolism, a potential topic for future research could be examining their relative contributions to particular effects. It may be that some associations are more dependent on articulatory features while others are more dependent on acoustic features. Of course, because acoustic and articulatory features are often inextricably linked (i.e., changes in articulation often result in changes in acoustics), this is an extremely difficult question to address. Even presenting linguistic stimuli visually for silent reading, or auditorily for passive listening, would not be sufficient to isolate acoustic features, as studies have shown that these can both lead to covert articulations (e.g., Fadiga, Craighero, Buccino, & Rizzolatti, 2002; Watkins, Strafella, & Paus,

2003). Moreover, as pointed out by Eberhardt (1940), even acoustic features such as frequency can have tactile properties (i.e., felt vibrations). Nevertheless, because mechanisms of association are often based on particular features (e.g., the statistical co-occurrence of acoustic frequency and size), pinpointing the features involved could help adjudicate between potential mechanisms’ roles in a certain effect. Future research might examine this by manipulating the component frequencies of vowels while maintaining their identity, or interfering with covert articulations, and then observing the effect on specific associations. In addition, beyond simply comparing the relative weighting of acoustic and articulatory features as a whole, it will be important to also consider the relative weighting among various acoustic features and articulatory features.

A related question is how individuals navigate the various associations afforded by phonemes’ bundle of features. For instance, what leads individuals to weigh certain phoneme features more heavily than others? Recall that the phoneme /u/ is associated with largeness

(Newman, 1933). This seems to suggest that individuals place more emphasis on the association afforded by this phoneme’s features as a back vowel (i.e., largeness) than as a high vowel (i.e.,

57 smallness).24 The matter is further complicated by the possibility that a given feature can afford associations with different ends of the same dimension. As an illustration, consider Diffolth’s

(1994) observation that the Bahnar language contains associations between high vowels and largeness (which contrasts with the typical Mil/Mal effect). Diffolth theorized that this resulted from a focus on the amount of space that the tongue takes up in the vocal tract (larger for high vowels), as opposed to the amount of space left empty (smaller for high vowels). Thus, this articulatory feature might potentially afford different (and conflicting) associations. Of course, this begs the question of why certain potential associations are more commonly observed than others (Nuckolls, 1999). Understanding what leads to the formation of certain associations out of the myriad of possibilities is an important topic for future research. This not only includes associations on an individual level, but also the crystallization of these associations in a given lexicon (in cases of indirect iconicity).

Lastly, it is worth briefly considering the role of visual features in sound symbolic effects. One example is the letters used to code for phonemes; visual features are sometimes presented as an important contributor to the Maluma/Takete effect (see Cuskley, Simner, &

Kirby, 2015). In fact, Cuskley et al. (2015) showed that the visual roundness/sharpness of letters was a stronger predictor of nonword-shape pairing than was consonant voicing. However, given that sound symbolic effects emerge in a culture without a writing system (Bremner et al., 2013), in preliterate infants (Ozturk et al., 2013; Peña et al., 2011; cf. Fort, Weiß, Martin, & Peperkamp,

2015; Pejovic & Molnar, 2016), with learned neutral orthographies (Hung et al., 2017), and are not affected by direct manipulations of font (Sidhu, Pexman, & Saint-Aubin, 2016), it seems probable that orthography is, as the very least, not the sole contributor to these effects.

24 Though this might be part of the reason why Newman (1933) found that /u/ was not rated as large as /ɔ/ (a mid- back vowel), for instance.

Nevertheless the contribution of orthography relative to those of acoustics and articulation is still an open question. If orthographic features were found to play a large role in sound symbolism, it might weaken the claims of some theories that rest on phonological and/or articulatory features

(e.g., the Frequency Code Hypothesis, double grasp neurons). Associations based in orthography would likely be due to shared low-level perceptual features among letters and associated stimuli

(though for potential roles of connotation, see Koriat & Levy, 1977; Walker, 2016). In addition, some articulatory features have very strong visual cues (e.g., lip rounding). It remains to be seen if it is possible to separate these features from the tactile properties of articulation.

Relationship with Crossmodal Correspondences

An open question is the extent to which discoveries regarding crossmodal correspondences can be applied to sound symbolism. In particular, one might wonder how well mechanisms of association for crossmodal correspondences can translate to sound symbolic associations. The issue is that while crossmodal correspondences involve simple, unidimensional stimuli (e.g., pure tones), linguistic stimuli are by their very nature more perceptually complex and multidimensional. Thus, while it might be tempting (and potentially correct) to explain sound symbolic associations as arising from a crossmodal correspondence of a phoneme’s component feature (e.g., between pitch and size; see Figure 5), the fact that the feature is embedded in a multidimensional stimulus will necessarily complicate matters (see Parise, 2016).

As an illustration of these complexities, D’Onofrio (2013) found that the influence of voicing in the Maluma/Takete effect was moderated by place of articulation. Nevertheless there is evidence that the two classes of effects are related. For instance, when Parise and Spence (2012) studied the Mil/Mal and Maluma/Takete effects using an IAT, they also examined crossmodal

59 correspondences (e.g., between pitch and size). All of these effects were found to have the same effect size, which was interpreted as being indicative of a common mechanism.

Figure 5. The relationship between sound symbolic associations and crossmodal correspondences (as the terms are used in this review). Sound symbolic associations are between a phoneme as a whole (including all of its component multidimensional features) and a particular stimulus dimension. Crossmodal correspondences are between simple stimulus dimensions (e.g., brightness or pitch). Crossmodal correspondences may exist between the component features of a phoneme and a particular dimension (illustrated by the example crossmodal correspondence between phoneme pitch and size), and could potentially contribute to a sound symbolic association of that phoneme.

A related question is the extent to which stimulus features are processed differently when they occur in linguistic vs. non-linguistic stimuli. For instance, does vowel pitch truly have the same associations as the pitch of a pure tone? Or does the involvement of the linguistic system alter the way that these associations operate (in addition to previously mentioned issues of multidimensionality)? Tsur (2006) theorized that linguistic stimuli could be processed based on their phonetic identity, their sensory features, or a combination of the two. It would stand to reason that an overlap between sound symbolic associations and crossmodal correspondences would depend on the stimuli being processed (at least in part) based on their sensory features.

Indeed there is some evidence for this. Fischer-Jørgensen (1968) found that Danish-speaking participants rated several pairs of allophones (e.g., [œ] and [ɑ], allophones of /æ/ as in had), differently on semantic differential scales. While allophones belong to the same phoneme category, they have different sensory features. Thus the fact that they were rated differently indicates that their sensory features affected their interpretation. We would not have expected this if they had been processed solely in terms of their phonetic identity.

Next Steps in Exploring Mechanisms of Association

Resolving the issues above will add to our understanding of sound symbolism. Still remaining, however, is the question of which of the proposed mechanisms underlie these associations. It is our opinion that the current body of experimental evidence does not allow us to definitively pinpoint any particular mechanism(s) as being responsible for sound symbolism.

This is in part because much of the existing work has focused on the effects of these associations, as opposed to the mechanisms that create them. Thus while a wealth of research exists, these experiments have not been designed to adjudicate between mechanisms of association.

It would be infeasible to test all five of the mechanisms at once, and so the best first step is likely to focus on developing testable hypotheses to adjudicate between two of them at a time.

In our opinion, this initial pair of mechanisms should be statistical co-occurrence and shared properties (particularly connotations) since these two mechanisms are best supported by the available empirical evidence. There is compelling evidence that statistical co-occurrences can create crossmodal correspondences between dimensions (e.g., Baier, Kleinschmidt, & Müller,

2006; Ernst, 2007; Teramoto, Hidaka, & Sugita, 2010; Zangenehpour & Zatorre, 2010), though this still remains to be demonstrated for sound symbolic associations. There is also evidence of similar connotations creating crossmodal correspondences (e.g., Walker, 2012; Walker &

Walker, 2012; Walker et al., 2012), and, more importantly, sound symbolic associations (e.g.,

Bozzi & Flores D’Arcais, 1967; Gallace et al., 2011; see also French, 1977). These results suggest these two mechanisms are the most promising starting points.

Using this pair of mechanisms as an example, future research aimed at adjudicating between mechanisms could take two tracks. One track would involve studies investigating whether each mechanism can create sound symbolic associations. This could be accomplished by attempting to create associations via either mechanism (i.e., artificially creating a statistical co-occurrence, and associating unrelated stimulus dimensions with some unifying shared property). One might also hypothesize as yet unmeasured sound symbolic effects based on either mechanism, and then test for those novel effects.

The other track would involve studies that examine existing sound symbolic associations in terms of whether they are better explained by statistical co-occurrence or a shared property. If a particular effect is due to a statistical co-occurrence then it should be possible to find evidence of that co-occurrence in the environment. In addition, we might expect manipulations of

62 individuals’ internalized probabilities to interfere with the association. If a particular effect is due to shared properties, then it should be possible to detect those shared properties via rating scales or reaction time measures. Another approach could be to examine if the strength of a given effect correlates with individual differences that would be relevant to a particular mechanism (e.g., differences in statistical learning; Misyak & Christiansen, 2012). Relatedly, sound symbolism effects have been shown to vary in some special populations (e.g., in individuals with Autism

Spectrum Disorder, Oberman & Ramachandran, 2008; Occelli, Esposito, Venuti, Arduino, &

Zampini, 2013; in individuals with dyslexia, Drijvers, Zaadnoordijk, & Dingemanse, 2015). To the extent that a given special population would be expected to differ in their capacity for a particular mechanism, this may represent another way of adjudicating between mechanisms.

Research that pits mechanisms against each other will be useful for generating evidence that some play a role in sound symbolism while others do not. While it is in principle possible that such research will discover that a single mechanism underlies all of sound symbolism, it seems more likely that multiple mechanisms contribute. The research reviewed in the preceding sections provides good reason to believe that a handful of mechanisms–even perhaps all those reviewed–play some role in sound symbolism. To the extent that this is borne out by future research, the next task for the field will be to examine the interplay between these mechanisms.

One possibility is that different mechanisms underlie different instances of sound symbolism. This suggests the intriguing possibility that certain mechanisms may be more likely to play a role for some kinds of dimensions than others. One potentially important distinction is that between prothetic (i.e., based on quantitative distinctions) and metathetic (i.e., based on qualitative distinctions) dimensions (Stevens, 1957). Gallace et al. (2011) hypothesized that for a metathetic domain such as taste, associations might be more likely to depend on shared

63 conceptual properties. Conversely, an account such as magnitude coding requires a prothetic domain. Another relevant factor might be the salience and/or prevalence of a given stimulus dimension, which could potentially affect the likelihood of statistical co-occurrence playing a role. One might also expect evolutionary factors be more influential for dimensions that are relevant to survival (e.g., size). Lastly, Ramachandran and Hubbard (2005) theorized that associations might be more likely to arise innately for stimulus dimensions that are represented in adjacent brain regions. Future research could compare mechanisms of association for dimensions that vary in these ways.

If it were demonstrated that different mechanisms underlie different effects, it would also be worthwhile for the field to consider if those different effects are indeed expressions of the same phenomenon. Perhaps it would be more accurate to view them as different kinds of sound symbolism–especially to the extent that they result in different behavioural effects. There is indeed some evidence of measurable differences between different instances of sound symbolism

(e.g., Vainio et al., 2016). A distinction that is often made in the crossmodal correspondence literature is between perceptual and decisional effects. The former involve genuine differences in perception (e.g., perceiving a dot as moving upwards when presented along with rising pitch;

Maeda, Kamai, & Shimojo, 2004) while the latter occur later in processing, and only involve effects on decisions, evident in reaction time or accuracy. Spence (2011) theorized that crossmodal correspondences arising from shared semantic features (in particular shared labels) would not lead to perceptual effects, while those based on co-occurrences or neural factors would lead to perceptual effects. Investigating the perceptual/decisional effect distinction across instances of sound symbolism could be productive. We may also expect associations arising from some mechanisms not to emerge on implicit measures. For instance, as speculated,

64 associations deriving from some shared properties may require explicit consideration. To the extent that associations with different origins lead to different behavioural outcomes, it may be prudent to consider them fundamentally different kinds of effects.

Another possibility is that multiple mechanisms combine to play a role in the same sound symbolic effect. For instance, it may be that the co-occurrence of two kinds of stimuli contributes to them having similar connotations. As noted, explanations based on shared properties beg the question of how stimuli come to be associated with those shared properties, and perhaps statistical co-occurrence could provide the answer in some instances.25 Conversely, it is possible that similar stimuli tend to co-occur more often (this is the basis of theories using lexical co-occurrence as a way of measuring meaning; e.g., Landauer & Dumais, 1997).

Magnitude coding represents another instance of mechanisms interacting (i.e., shared properties and neural factors). In this case, stimuli from different dimensions have the shared property of high (or low) magnitude, but the association fundamentally results from the neural coding of that property.

This interplay between mechanisms seems especially relevant to evolution-based theories. Consider the fact that Ohala’s (1994) Frequency Code Hypothesis involves an evolved sensitivity to a statistical co-occurrence. This presents the intriguing possibility that while some statistical co-occurrence based associations must be learned, others have become innate via evolutionary processes. Note, however, that Ohala (1994) concedes that some postnatal experience may be required in the formation of the Frequency Code. Thus, perhaps it would be more correct to say that there is an evolved predisposition to acquire associations based on

25 However, statistical co-occurrence would certainly not apply in every instance. As Walker and Walker (2012) point out, though small and bright objects share connotations, we would not expect smallness to co-occur with surface brightness.

65 certain statistical co-occurrences. In particular, this seems more likely to apply to co-occurrences that are based on fundamental physical laws, rather than those that may vary locally. Similarly, evolved predispositions may play a role in some phonemes and stimuli sharing affective properties (Nielsen & Rendall, 2011; 2013). In our review we have treated each of the five mechanisms as distinct, but there are many ways in which they could interact in the production of sound symbolism. Moreover, some mechanisms may be so interdependent that they cannot be understood in isolation (e.g., shared properties arising via co-occurrence).

As the preceding examples illustrate, while multiple mechanisms may play a role in a single effect, they need not do so simultaneously. On the contrary, several mechanisms may play out sequentially in the creation of an effect. This could be true in terms of both ontogeny and phylogeny. In addition, when considering the contribution of multiple mechanisms to an observed behavioural effect, some may be more proximally related to that effect than others. As an illustration, consider an instance in which statistical co-occurrence leads to stimuli sharing a connotation; while both mechanisms would contribute to an observed behavioural effect, the stimuli sharing a connotation may do so more proximally. Of course, it is also possible that in some effects, phonemes are simultaneously associated with stimuli by multiple separate mechanisms of association that do not interact (see D’Onofrio, 2013; Nichols, 1971). A major challenge for the field going forward will be untangling these complex interactions.

Conclusion

Sound symbolism refers to an association between phonemes and particular kinds of stimuli. It provides a means by which language can be non-arbitrary, by facilitating iconic relationships between form and meaning. A variety of mechanisms have been proposed to explain how acoustic and articulatory properties of phonemes come to be associated with other

66 stimuli. The associations may arise due to phoneme features and related stimuli co-occurring in the world. Another possibility is that phoneme features and associated stimuli share a common property, be it perceptual, conceptual, affective or linguistic. The associations may also be due to structural properties of the brain, evolution, or patterns extracted from language. While there is a wealth of experimental evidence on the effects of sound symbolic associations, there has been much less work on the mechanisms that might create them. It is our hope that the preceding review will foster such investigations. We suggest that future investigations should be focused around the following points:

(a) Examining whether each of the different mechanisms can and do contribute to sound

symbolic associations, potentially beginning with further investigation into the

mechanisms of statistical co-occurrence and shared properties.

(b) If evidence suggests that different mechanisms underlie different associations, examining

whether some mechanisms are more likely for particular kinds of dimensions than others,

and if associations created by different mechanisms result in different behavioural effects.

effect, examining the interplay of those contributions.

The study of sound symbolism reveals hidden dimensions of richness and meaning in language. For instance, Jorge-Luis Borges (1980) opined “…the English [word] moon has something slow, something that imposes on the voice a slowness that suits the moon” (p. 62).

We might speculate that this arises from the association between nasal sonorants (e.g., /m/ and

/n/) and back vowels, and slowness (Cuskley, 2013; Saji et al., 2011). Such sound symbolic associations illuminate the multimodal nature of human cognition. As interest in sound symbolism increases, the focus of future research must shift to understanding the mechanisms

67 that underlie such associations. The field must test predictions derived from extant theories, and work to refine those theories. We have offered some ideas for that future work here, and are confident that the years to come will bring with them a fuller and deeper understanding of this fascinating phenomenon.

Acknowledgements

This research was supported by the Natural Sciences and Engineering Research Council of

Canada (NSERC) through a postgraduate scholarship to DMS and a Discovery Grant to PMP; and by Alberta Innovates: Health Solutions (AIHS) through a graduate scholarship to DMS.

We would like to thank: our two reviewers for their very helpful suggestions; Suzanne Curtin for her helpful comments on an earlier version of this manuscript; Michele Wellsby and Lenka

Zdrazilova for their helpful comments on a draft of this manuscript; Paolo Flores D’Arcais,

Charles Hutton, John Joseph, Alan Nielsen, Luca Nobile, Cesare Parise, and Peter Walker for their helpful replies to email queries; Alberto Umiltà for providing Italian translation; and

Padraic Monaghan for providing systematicity values.

Chapter 3: Investigating Phoneme-Personality Sound Symbolism in First Names

Copyright © 2019 by American Psychological Association. Reproduced with permission. Sidhu, D. M., Deschamps, K., Bourdage, J. S., & Pexman, P. M. (2019). Does the name say it all? Investigating phoneme-personality sound symbolism in first names. Journal of Experimental Psychology: General. doi:10.1037/xge0000662 No further reproduction or distribution is permitted without written permission from the American Psychological Association.

Introduction

Sound Symbolism

Sound symbolism is the phenomenon by which certain phonemes seem inherently associated with certain kinds of things (for recent reviews, see Lockwood & Dingemanse, 2015;

Sidhu & Pexman, 2018a). As an example, consider the nonwords maluma and takete, and the round and sharp shapes shown in Figure 6. When asked which of the nonwords go with which of the shapes, approximately 90% of people (see Styles & Gawne, 2017) pair maluma with the round shape and takete with the jagged shape. Something in the sound and/or articulation of these nonwords leads to the sense that they go along better with certain kinds of shapes; this is a sound symbolic association. In addition to such overt effects, sound symbolism also emerges implicitly in that individuals have a faster reaction time when responding to sound symbolically congruent vs. incongruent pairings of stimuli (e.g., maluma with a round vs. a jagged shape; e.g.,

Hung, Styles, & Hsieh, 2017; Ohtake & Haryu, 2013). There are also differences in event related potentials when processing congruent vs. incongruent sound-shape pairings, suggesting that these congruencies have effects beyond observable behaviour (e.g., Asano et al., 2015; Ković,

Plunkett, & Westerman, 2010; Sučevič, Savić, Popović, Styles, & Ković, 2015).

Figure 6. An illustration of the maluma/takete effect, in which most individuals judge the nonword maluma as a better match for the round shape on the left, and the nonword takete as a better match for the sharp shape on the right.

The example given in Figure 6 is one of the most well studied sound symbolic effects, and has come to be known as the maluma/takete effect (Köhler, 1929). The effect goes beyond this pair of nonwords and applies to groups of phonemes with similar acoustic and articulatory properties. In general, certain sonorant consonants (/m/, /n/ and /l/; see Table 4 for definitions of linguistic terms), voiced stop consonants (/b/, /d/ and /g/, though to a lesser extent; cf. Bottini,

Barilari, & Collignon, 2019), and back rounded vowels (e.g., /u/ as in who’d, and /oʊ/ as in hoed) show an association with round shapes; while voiceless stop consonants (e.g., /p/, /t/ and /k/), and high-front unrounded vowels (e.g., /i/ as in heed) show an association with sharp shapes (e.g.,

D’Onofrio, 2013; McCormick, Kim, List, & Nygaard, 2015; Nielsen & Rendall, 2011). The maluma/takete effect is not the only example of sound symbolism. Another is the mil/mal effect: an association between high-front vowels (e.g., /i/) and small shapes, and between low-back

71 vowels (e.g., /ɑ/, as in hawed) and large shapes (Newman, 1933; Sapir, 1929). Associations have also been demonstrated between certain phonemes and other perceptual dimensions, such as brightness (e.g., Newman, 1933), speed (Cuskley, 2013), hue (Moos, Smith, Miller, & Simmons,

2014), and taste (Gallace, Boschin, & Spence, 2011).

Table 4

Definitions of linguistic terms used throughout the article (derived from Ladefoged & Johnson,

2010; Reetz & Jongman, 2009).

Phoneme Term Examples

Back vowels are those articulated with the /u/ as in who’d, /ɑ/ as in hawed highest point of the tongue relatively close to the back of the mouth.

Bilabial consonants involve the lips coming /m/ as in mat, /b/ as in bat together in their articulation.

Front vowels are those articulated with the /i/ as in heed, /æ/ as in had highest point of the tongue relatively close to the front of the mouth.

High vowels are those articulated with the /i/ as in heed, /u/ as in who’d tongue relatively close to the roof of the mouth.

Low vowels are those articulated with the /æ/ as in had, /ɑ/ as in hawed tongue relatively far from the roof of the mouth.

Rounded vowels are those articulated with /u/ as in who’d, /oʊ/ as in hoed rounded lips.

Sonorant consonants involve no stoppage of, /m/ as in mac, /l/ as in lack or turbulence in, the airflow; this includes nasals and approximants.

Stop consonants involve a stoppage of /p/ as in pat, /b/ as in bat airflow.

Unrounded vowels are those articulated /i/ as in heed, /æ/ as in had without rounded lips.

Voiced consonants involve the vocal folds /b/ as in bam, /d/ as in dam being brought close enough together to vibrate.

Voiceless consonants involve the vocal folds /p/ as in pat, /t/ as in tat not being brought close enough together to vibrate

In the five studies to be presented here we conducted the first systematic investigation of a different form of sound symbolism: associations between phonemes in real first names and personality factors. This new type of sound symbolism is relevant to several important topics and theoretical issues. First, these studies address the extent to which sound symbolism can emerge in the context of existing language. This is an important question as it has broad implications for the relevance of sound symbolism to language processing. Second, this work explores sound symbolic associations for an abstract dimension (i.e., personality), in contrast to the perceptual

73 dimensions that are typically studied. This broadens the scope of sound symbolism and serves as an informative test case for the mechanisms that can give rise to it. Finally, the information gleaned from a name via its sound symbolic associations has relevance for the study of impression formation, as a name is often one of the first pieces of information we receive about a person.

Sound Symbolism in Real Language

Sound symbolic associations are relevant to the fundamental nature of language; in particular, to the relationship between the form of a word (i.e., its phonology, articulation and orthography) and its meaning. One position is that the relationship is arbitrary, and that the form of a word does not have any kind of special relationship to its meaning (e.g., Hockett, 1963).

Consider the word fun, for instance: there is nothing particularly “fun” about it. On the contrary, it seems to be comprised of an arbitrarily chosen set of phonemes. However, sound symbolic associations provide one avenue by which a word’s form–in particular its phonology–can be related to its meaning. For instance, the word balloon contains phonemes that are sound symbolically related to roundness; it also refers to a round object. Thus, balloon is an example of non-arbitrariness in language.

Instead of conceptualizing arbitrariness and non-arbitrariness as mutually exclusive categories, there has been a recent shift towards viewing them as complementary aspects of language (Dingemanse, Blasi, Lupyan, Christiansen, & Monaghan, 2015; Perniss, Thompson, &

Vigliocco, 2010). That is, words can contain both arbitrary and non-arbitrary aspects. For instance, while balloon contains phonemes related to roundness (a non-arbitrary property) it is also to some extent arbitrary–there is no reason that the round-associated phonemes in balloon had to be arranged in that order. Its partial arbitrariness is also highlighted by the fact that it

74 would be highly unlikely for someone to guess balloon’s exact meaning based on its form alone.

By this view, non-arbitrariness is present in different amounts throughout the lexicon, and is a general property of language (see Perniss et al., 2010; Perry, Perlman, & Lupyan, 2015). Because sound symbolism contributes to non-arbitrariness, understanding the phenomenon of sound symbolism is a key question for the study of language.

Despite this potential relevance to real language, sound symbolism has largely been examined using nonwords and the extent to which sound symbolism has an effect in real words

(i.e., whether the round-associated phonemes in balloon affect its processing) is still unclear.

Examinations of the maluma/takete effect in existing language have been equivocal (Sidhu &

Pexman, 2015; Sučević, Janković, & Ković, 2013; Sučević, Savić, Popović, Styles, & Ković,

2015; Westbury, 2005). Indeed some models of word processing lead to the prediction that sound symbolism effects should be attenuated with real language as compared with nonwords. For instance, the Dual Route Cascaded model (DRC; Coltheart, Rastle, Perry, Langdon, & Ziegler,

2001) of word recognition predicts less extensive phonological processing for real words as compared to nonwords. It also predicts a greater amount of semantic activation for real words, which might interfere with the perceptual/semantic features activated via phonology (e.g., sonorants sound symbolically activating roundness). Testing sound symbolism in real words (in the form of first names) was one of the main goals of this paper.

Phoneme-Personality Sound Symbolism

Another goal of this paper was to examine sound symbolism beyond the concrete perceptual dimensions in which it is typically studied. One relevant question is whether sound symbolism also exists between phonemes and more abstract targets (i.e., those that are not tangible, and cannot be experienced with the senses; see Brysbaert, Warriner, & Kuperman,

2014). This is especially challenging given that abstract concepts lack perceptual features that would invite obvious comparisons to the perceptual properties of phonemes. Indeed, there is a large literature concerned with how abstract concepts are learned and represented, given that they do not have obvious ties to sensorimotor experience (for a review see Borghi et al., 2017). Thus, it is meaningful to examine the extent to which sound symbolism will extend to abstract targets.

This can have implications for our understanding of the mechanisms that can enable sound symbolic associations.

There is a rich literature examining the semantic connotations of various phonemes from the mid to late 20th century (e.g., Bozzi & Flores D’Arcais, 1963; Greenberg & Jenkins, 1966;

Miron, 1961; Tarte, 1982). While some of these are indeed abstract in nature (e.g., pleasant- unpleasant) others are related to perceptual dimensions (e.g., abrupt-continuous). Recent work has shown an association between certain phonemes and the abstract dimensions of social dominance (Auracher, 2017) and valence (Rummer, Schweppe, Schlegelmilch, & Grice, 2014).

Note that both of these involve perceptual features, in the posture stimuli used, and the hypothesized mediator of facial expression, respectively.26

In the experiments described here, we investigated whether phonemes show a sound symbolic association with the abstract construct of personality. Personality refers to individual differences in patterns of thinking, feeling, and behaving. Language plays a critical role in our understanding of personality. Indeed, according to the lexical hypothesis, individual differences become encoded in human language (Allport & Odbert, 1936; Galton, 1884), and by factor

26 Other studies showing abstract associations of phonemes include: preferences for labels whose phonemes are articulated with an approach vs. avoidance sequence (Topolinksi, Maschmann, Pecher & Winkielman, 2014), and the demonstration that high-front vs. low-back phonemes lead to more precise construals (Maglio et al., 2014). We do not mention these in the main text because they involve a sequence of, rather than individual, phonemes; and an association with cognitive styles rather than dimensions; respectively.

76 analyzing responses to words and determining which sets of words cluster together, researchers have been able to determine the broad dimensions and structure of personality (e.g., Ashton et al., 2004; Goldberg, 1990; Tupes & Christal, 1992). Personality, particularly at the factor level, represents an ideal domain in which to test abstract sound symbolism. While individual traits may have salient perceptual features, personality factors are latent constructs, and thus relatively abstract.

An ideal way to investigate this would be in the context of first names, for two main reasons. First, using real first names would be a way of examining whether sound symbolism emerges in the context of real language. As mentioned earlier, there are reasons to believe that sound symbolism effects will be attenuated when using real words. While names may not have a meaning per se, they are presumed to activate identity-specific semantics, which in turn activate general semantics (see Valentine, Brennen, & Brédart, 1996). For instance, hearing the name

Bob might bring to mind one’s friend Bob who is a zookeeper, and could thus also activate semantics related to that profession. Thus, names allow for investigation of whether sound symbolism effects will emerge in the presence of some existing semantic information. Second, there is evidence that individuals will make inferences based on the phonology of a name. For instance, Cassidy, Kelly and Sharoni (1999) found that individuals are quicker to classify names as female or male if they contain gender consistent phonology (e.g., female names that end in a schwa, as in Erica; see also Slepian & Galinsky, 2016). Note that this was not due to sound symbolism per se, but rather the distributional properties of phonemes in the names. However, there is a good deal of evidence that the sound symbolic properties of invented product names can impact perceptions of products. For instance, Lowrey and Shrum (2007) demonstrated that individuals prefer products with names that have sound symbolic associations that are desirable

77 for that product (e.g., sharpness for knives; see also Klink, 2000). Velasco, Salgado-Montejo,

Marmolejo-Ramos and Spence (2014) have also demonstrated that the sound symbolism of product names can impact expectations of taste (for a review of this area see Spence, 2012).

A link between names and personality also has implications for our understanding of impression formation. Researchers have had a keen interest in increasing our understanding of how people form impressions of others. A number of factors have been examined in that regard, including but not limited to facial features (e.g., Vernon, Sutherland, Young, & Hartley, 2014), ethnicity (e.g., Cottrell & Neuberg, 2005) and gender (e.g., Oh, Buck, & Todorov, 2019).

Another area that has received interest is how names guide these impressions, as a name is one of the first things we learn about somebody in various contexts, including a job application, a blind date, or an e-introduction. For instance, names have been investigated as predictors of perceptions of intelligence and competence (Young, Kennedy, Newhouse, Browne, & Thiessen,

1993), teacher expectations of a student’s achievement motivation (Anderson-Clark, Green, &

Henley, 2008), and resume call-backs in the job search (Bertrand & Mullainathan, 2004). In looking at why names can have an impact on these perceptions, research has examined characteristics of a name such as: ethnicity (Bertrand & Mullainathan, 2004), formality (Leirer,

Hamilton, & Carpenter, 1982), and pronounceability (Laham, Koval, & Alter, 2012). In a series of studies that are close to the work at hand, Mehrabian and Piercy found that names differ in their semantic connotations, and that this is driven in part by their length (1993a) and spelling conventionality (1993b). The present set of studies expanded on this work by examining another source of information in names: the sound symbolic associations of the phonemes that they contain.

While much of the work on sound symbolism has involved associations with perceptual dimensions, there have been some explorations of links between phonemes and particular personality traits. Milán, Córdoba, Juárez-Ramos, Artacho and Rubio (2013) found that, compared to bouba, kiki was happy, clever, unpleasant and nervous. Shinohara and Kawahara

(2013; see also Kawahara, Shinohara & Grady, 2015) found that sonorants were associated with cute, soft and accessible personalities, while obstruents were associated with sharp, inaccessible and blunt personalities.

The studies by Milan et al., and Shinohara and Kawahara used invented words. This reduces the generalizability of these findings because, as discussed previously, there is reason to believe that sound symbolism might operate differently in real words and nonwords. Sidhu and

Pexman (2015) found preliminary evidence that real first names may be associated with certain personality traits based on their phonology. They asked participants to generate adjectives that they would associate with someone with a “round and curvy personality” or a “sharp and spiky personality”. The adjectives participants generated included: easygoing and friendly, and determined and rigid, respectively. They then presented a separate group of participants with pairs of gender matched names, one of which contained consonant phonemes typically associated with round shapes (e.g., Molly or Noel), and one of which contained consonant phonemes typically associated with sharp shapes (e.g., Kate or Kirk). Participants were asked which name was more likely to belong to a person possessing one of the previously generated traits. Participants were more likely to indicate that names like Molly belonged to people possessing the “round” personality traits, while names like Kate belonged to people possessing the “sharp” personality traits. However, one might argue that this approach “stacked the deck” in

79 favour of finding an association; in particular, by choosing traits that were most strongly associated with the concepts of roundness and sharpness (see Table 5).

Table 5

“Round” and “Sharp” traits generated by participants and used as stimuli by Sidhu and Pexman

(2015) in their Experiment 2.

“Round” Traits “Sharp” Traits

Adaptable Aggressive

Easygoing Angry

Friendly Determined

Funny Harsh

Introverted Irritable

Nice Jumpy

Open Mean

Sensitive Rigid

Unreliable Sarcastic

Versatile Unfriendly

Taken together, there has been some evidence of sound symbolic associations with personality. However, this has been produced in a piecemeal fashion, with each study testing a few specific personality descriptors. In addition, studies have explored traits for which researchers have an a priori reason to expect an association with sonorants or voiceless stops.

Thus, there has yet to be a comprehensive and unbiased exploration of sound symbolism and

80 personality space in its entirety. The limited approach of previous work also makes it impossible to make any claims regarding associations between phonemes and higher order factors of personality.

The Present Study

In the present study we addressed these shortcomings with a thorough exploration of personality space, and one that allowed us to examine relationships with higher order personality factors. We used a full sampling of traits from the six factors of the HEXACO model of personality (Lee & Ashton, 2004). The HEXACO model of personality is a six factor framework of personality that has been derived from lexical studies (using similar or the same adjective sets to those used to derive the Big Five) and replicated across more than a dozen languages (e.g.,

Ashton et al., 2004; Lee & Ashton, 2008). This model contains six personality factors: Honesty-

Humility, Emotionality, Extraversion, Agreeableness, Conscientiousness, and Openness to

Experience (see Table 6 for a description of each). In comparison to the Big Five, HEXACO

Conscientiousness, Openness, and Extraversion factors correspond strongly with their Big Five counterparts. Conversely, Emotionality and Agreeableness are rotational variants of the Big Five factors of Neuroticism and Agreeableness, with HEXACO Emotionality containing content related to sentimentality and sensitivity, and HEXACO Agreeableness containing content related to anger and hostility (Lee & Ashton, 2004). The largest difference between the HEXACO and

Big Five models is the presence of a sixth factor of personality, namely Honesty-Humility. Note that numerous recent studies have supported a six- vs. a five-factor model of personality (e.g.,

Ashton et al., 2004; Ashton, Lee, & Goldberg, 2004; Lee & Ashton, 2008). To date, the

HEXACO model has been used in hundreds of studies (see hexaco.org/references) and possesses

81 a number of practical and theoretical advantages over and above the Big Five (see Ashton & Lee,

2007, for a review).

Table 6

A description of each factor of the HEXACO, taken verbatim from Lee and Ashton (2009).

Personality Factor and Description

Honesty-Humility: Persons with very high scores on the Honesty-Humility scale avoid manipulating others for personal gain, feel little temptation to break rules, are uninterested in lavish wealth and luxuries, and feel no special entitlement to elevated social status.

Conversely, persons with very low scores on this scale will flatter others to get what they want, are inclined to break rules for personal profit, are motivated by material gain, and feel a strong sense of self-importance.

Emotionality: Persons with very high scores on the Emotionality scale experience fear of physical dangers, experience anxiety in response to life's stresses, feel a need for emotional support from others, and feel empathy and sentimental attachments with others. Conversely, persons with very low scores on this scale are not deterred by the prospect of physical harm, feel little worry even in stressful situations, have little need to share their concerns with others, and feel emotionally detached from others. eXtraversion: Persons with very high scores on the Extraversion scale feel positively about themselves, feel confident when leading or addressing groups of people, enjoy social gatherings and interactions, and experience positive feelings of enthusiasm and energy.

Conversely, persons with very low scores on this scale consider themselves unpopular, feel

82 awkward when they are the center of social attention, are indifferent to social activities, and feel less lively and optimistic than others do.

Agreeableness: Persons with very high scores on the Agreeableness scale forgive the wrongs that they suffered, are lenient in judging others, are willing to compromise and cooperate with others, and can easily control their temper. Conversely, persons with very low scores on this scale hold grudges against those who have harmed them, are rather critical of others' shortcomings, are stubborn in defending their point of view, and feel anger readily in response to mistreatment.

Conscientiousness: Persons with very high scores on the Conscientiousness scale organize their time and their physical surroundings, work in a disciplined way toward their goals, strive for accuracy and perfection in their tasks, and deliberate carefully when making decisions.

Conversely, persons with very low scores on this scale tend to be unconcerned with orderly surroundings or schedules, avoid difficult tasks or challenging goals, are satisfied with work that contains some errors, and make decisions on impulse or with little reflection.

Openness: Persons with very high scores on the Openness to Experience scale become absorbed in the beauty of art and nature, are inquisitive about various domains of knowledge, use their imagination freely in everyday life, and take an interest in unusual ideas or people.

Conversely, persons with very low scores on this scale are rather unimpressed by most works of art, feel little intellectual curiosity, avoid creative pursuits, and feel little attraction toward ideas that may seem radical or unconventional.

We examined whether names containing sonorants vs. voiceless stops were differentially associated with any of the factors from the HEXACO, in laboratory tasks (Experiments 1, 2, 4

83 and 5) and in the self-reported personalities of a large adult sample (Experiment 3). We examined these two groups of phonemes (i.e., sonorants vs. voiceless stops) because they have been shown to have distinct sound symbolic associations (e.g., in the maluma/takete effect).

While vowels also contribute to sound symbolism, the limitation of using existing names precluded a fully balanced design manipulating both consonants and vowels.

Experiment 1

In Experiment 1 we tested for phoneme-personality sound symbolism by examining whether participants were more likely to choose a name containing sonorants vs. voiceless stops as belonging to someone who possessed specific personality traits. Together, these traits represented the six factors of the HEXACO.

Method

Ethics Statement. All experiments reported in this paper were approved by the Conjoint

Faculties Research Ethics Board at the University of Calgary, and were carried out in accordance with the provisions of the World Medical Association Declaration of Helsinki. All participants gave written (Experiments 1, 2, and 4) or online (Experiments 3 and 5) consent, and were debriefed after the experiment.

Participants. Participants were 60 undergraduate students (47 female; M age = 20.75; SD

= 3.68) at the University of Calgary who participated in exchange for course credit. This number was nearly double the 32 participants examined by Sidhu and Pexman (2015; Experiment 2). We used the “simr” package (Green & Macleod, 2016) in the statistical software R (R Core Team,

2016) to conduct a power analysis using the data from Sidhu and Pexman (2015), as this is the closest existing experiment to the one we had planned. Using simulation, we determined that a sample size of 60 participants would have had a power of 0.999 to detect an effect of adjective

84 type (i.e., round- vs. sharp-associated), of the size that was observed in those data when random intercepts are included (b = 0.59), with α = .05. Thus we were confident that the present experiment had enough power to detect an effect. All participants reported English fluency and normal or corrected to normal vision.

Materials and Procedure. Trait stimuli were selected based on the loadings of 449 traits onto the six factors of the HEXACO (Lee & Ashton, 2008). For each factor, we chose the three traits that loaded most heavily onto its high and its low ends (i.e., traits indicative of being high or low in that particular personality factor), with the following exceptions. We did not use any traits with multiple loadings (e.g., warm-hearted which loaded onto both Agreeableness and

Extraversion). We also did not use traits that referred directly to gender (i.e., masculine/feminine), that involved the same compound (e.g., quick-tempered excluded hot- tempered) or that could be interpreted as a political affiliation (i.e., conservative). See Table B1 in Appendix B for a list of traits.

Name stimuli were 72 first names (36 male; 36 female) from The Alberta Services list of registered baby names in Alberta in 2014. These norms are available at http://www.servicealberta.ca/Alberta_Top_Babies_Names.cfm. We selected two groups of names that contained consonants with opposing sound symbolic associations (e.g., as in the maluma/takete effect; Nielsen & Rendall, 2011, 2013) and connotative-semantic properties (e.g.,

Greenberg & Jenkins, 1966). In particular, we chose the sonorants /m/, /n/ and /l/ to constitute one group, and the voiceless stops /p/, /t/ and /k/ to constitute the other. Half of the names of each gender contained at least one of these sonorants27 and no voiceless stops (i.e., sonorant names); the other half contained at least one voiceless stop and none of these sonorants (i.e.,

27From this point forward we use the term sonorants to refer specifically to those implicated in the maluma/takete effect (i.e., /m/, /n/, and /l/).

85 voiceless stop names). Balancing gender of name types was important, as females and males have been found to differ on several personality factors (Lee & Ashton, 2004). The names were presented in same-gender pairs consisting of one name of each type (i.e., one sonorant name and one voiceless stop name). The names in each pair were matched according to frequency, length within one syllable, and location of the sound symbolic phoneme (e.g., a name beginning with a sonorant would be paired with a name beginning with a voiceless stop). We attempted to choose names that included sound symbolically congruent vowels (i.e., sonorant names with back vowels; voiceless stop names with front vowels). However, because we were constrained by existing names, and the matching efforts mentioned previously, we weren’t able to do this perfectly. Nevertheless, sonorant names contained a significantly higher percentage of back vowels (M = 38.43; SD = 34.00) than did voiceless stop names (M = 2.78; SD = 11.62), t(70) =

5.95, p < .001. In addition, sonorant names contained a significantly lower percentage of front vowels (M = 20.37; SD = 29.84) than did voiceless stop names (M = 67.59; SD = 37.57), t(70) =

5.91, p < .001. See Table B2 in Appendix B for name pairs and their frequencies.

Each trial began with the presentation of a fixation cross for 1,000 msec, followed by a blank screen for 500 msec. A trait was then presented in the center of the screen with two paired names presented in the bottom left- and right-hand corners of the screen. Participants were instructed to think of these names as people they had never met and to indicate, based on the names, who they thought the adjective would best describe. Responses were made via button press. The names remained on the screen until a response was given, after which a blank screen was presented for 500 msec between trials. Participants saw each name pair once for a total of 36 trials. The pairings of names and traits, the order in which they were presented, and the side of the screen on which a given name appeared, were all randomized and counterbalanced across

86 participants. Next, in a separate task, all participants were asked to rate the familiarity of each presented name on a scale from 1 (“not familiar at all”) to 5 (“extremely familiar”). Participants were given the following examples of individuals whose names would likely be extremely familiar: a family member, a partner, a best friend, or a favourite TV character.

Results

Data for all experiments, and code for all analyses, can be found at the following online data repository: https://osf.io/r84s2/. The data of in lab experiments were analyzed using mixed effects regressions. Following the suggestions of Barr, Levy, Scheepers and Tily (2013), analyses included all possible random effects. That is, models always included random subject and item (i.e., presented name[s]) intercepts, as well as random subject and/or item slopes (where appropriate) for any fixed effects included in a model. Models in all experiments were fit using

Bayesian parameter estimation. In short, this approach determines the probability that a model’s parameters take on different values, given the observed data (viz., the posterior). Following

Bayes’ theorem, this is proportional to a combination of prior expectations for those parameter values (viz., the prior) and the likelihood that we would have observed our data given different parameter values (viz., the likelihood). In practice, functions describing the prior and the likelihood are combined to create a posterior density function. This is then sampled from28, and the resulting distribution can be used to establish an estimate of the parameter, and its 95% credible interval: the range of values with a 95% probability of containing the true value of a given parameter. When a given parameter’s credible interval does not include zero, this is considered sufficient evidence for that parameter having a statistically credible effect on the outcome measure.

28In particular, this method used a No-U-Turn Sampler (Hoffman & Gelman, 2014).

All models were computed using the statistical software R (R Core Team, 2016) and the package “brms” (Bürkner, 2017), which fits Bayesian mixed effects models using the Stan programming language. Analyses were run using 20 sampling chains, each with 2,000 iterations; the first 1,000 of these were treated as warmups, resulting in 20,000 posterior samples. In cases where the parameters of interest had an effective sample size lower than 10,000 (Kruschke,

2015), the model was rerun with additional chains until this threshold was met. Because of the lack of previous literature on the topic, models were fit using a generally accepted weakly informative prior for fixed effects (i.e., a Cauchy distribution centered at zero, with a scale of two and a half) and intercepts (i.e., a Cauchy distribution centered at zero, with a scale of ten;

Gelman, Jakulin, Pittau, & Su, 2009). In order to implement these priors, all continuous predictors and outcome variables were scaled to have a mean of zero and a standard deviation of

0.50; all binary predictors were effects coded and (when necessary; i.e., when there were unequal numbers in each group and thus the predictor did not already have a mean of zero) shifted to have a mean of zero, and to differ by one between their two values (Gelman et al., 2009). We used the package default weakly informative prior for random effects (i.e., half Student’s t distribution with three degrees of freedom). All R̂ values were ≤ 1.01, indicating that the analysis had reached convergence (i.e., additional sampling would not have lead to different results;

Gelman & Rubin, 1992).

In the present experiment, we analyzed trials presenting traits from each of the six personality factors separately using mixed effects logistic regressions, and interpreted the models’ intercepts. Our dependent variable was whether participants chose the sonorant vs. the voiceless stop name. On trials presenting traits from the high end of a factor, selecting the sonorant name was coded as “1” while selecting the voiceless stop name was coded as “0”. This

88 coding was reversed for trials presenting traits from the low end of a factor. Thus, the intercepts of these models reflected the combined likelihood of participants choosing the sonorant name as being higher in a given factor, and choosing the voiceless stop name as being lower in a given factor. Our results indicated that participants were more likely to select the sonorant (voiceless stop) name as being higher (lower) in Emotionality [95% credible interval, presented for untransformed data; 0.10, 0.67], Agreeableness [0.34, 0.98] and Conscientiousness [0.09, 0.58].

In particular, the odds were 1.46 times higher for Emotionality, 1.92 times higher for

Agreeableness, and 1.41 times higher for Conscientiousness (see Table 7).

Table 7

Resulting intercepts of logistic regressions predicting the likelihood of selecting a sonorant

(voiceless stop) name for traits from the high (low) end of each personality factor, in Experiment

Factor Intercept SE 95% CIs

Honesty-Humility 0.26 0.15 [-0.02, 0.56]

Emotionality 0.38 0.14 [0.10, 0.67]*

Extraversion -0.13 0.19 [-0.52, 0.26]

Agreeableness 0.65 0.16 [0.34, 0.98]*

Conscientiousness 0.34 0.12 [0.09, 0.58]*

Openness 0.18 0.15 [-0.10, 0.48]

* 95% Credible Interval does not include zero.

In a set of supplementary analyses we examined if participants’ reported familiarity with the names, and/or the gender of the names presented on a given trial (effects coded; female names = 0.5; male names = -0.5), affected the likelihood of choosing one name over the other.

Each of the previous models was rerun with these predictors, as well as their interaction, included. A familiarity score was computed for each trial by taking the difference in familiarity between the name whose selection would be coded “1” (e.g., the sonorant name on trials involving a high end trait) and whose selection would be coded “0” (e.g., the voiceless stop name on trials involving a high end trait). Both measures had a mean of zero in trials for each personality factor, allowing us to examine the effect of their inclusion on the models’ intercepts.

Emotionality [0.11, 0.76], Agreeableness [0.46, 1.52], and Conscientiousness [0.14, 0.69] still had intercepts whose credible intervals did not include zero. Familiarity, gender, and their interaction, were not statistically credible predictors in any of the models.

Discussion

The results of Experiment 1 demonstrated that names containing sonorant vs. voiceless stop phonemes were more likely to be judged as belonging to someone who is high in

Emotionality, Agreeableness, and Conscientiousness. This was true even after adding name familiarity and name gender to the models.

There are several drawbacks to the approach taken in this experiment. One is that the nature of the task forced participants to consider the relationship between a pair of names. This could serve to highlight relevant differences in their phonology. It could be that certain names only seem like good (or bad) matches for a certain trait when considered in relation to a name with contrasting phonology. The forced choice task was also a rather insensitive measure, unable to capture variations in the extent to which a given name did or did not seem to go with a given

90 trait. In addition, only being able to examine the effect of trial gender on the model’s intercept was a rather inelegant way of accounting for the effects of name gender. We addressed these issues in Experiment 2 by presenting participants with a single name on each trial and asking participants to make a continuous rating of the fit of that name with a trait. This allowed us to get a precise measure of each type of name’s fit with personality factors in isolation, and to model the interaction between name type and name gender.

Experiment 2

In Experiment 2, we tested for phoneme-personality sound symbolism by asking participants to rate how well they thought someone with a certain name would be described by a given trait.

Method

Participants. Participants were 60 undergraduate students (45 female; M age = 21.63;

SD = 4.61) at the University of Calgary who participated in exchange for course credit. As this study measured the same construct as Experiment 1 (albeit with a continuous rather than dichotomous outcome variable), we chose to test the same number of participants as in that experiment. Note that one participant’s age was not recorded. All participants reported English fluency and normal or corrected to normal vision.

Materials and Procedure. The name stimuli in Experiment 2 were the same as those used in Experiment 1, except here they were presented one at a time. In addition, Experiment 2 employed a Likert scale rating as opposed to a forced choice task. Each trial began with a trait presented on screen for 2,000 msec, followed by a blank screen for 500 msec. A single name then appeared in the center of the screen with the rating scale below it from 1 (“not at all”) to 7

(“extremely”). Participants were instructed to think of the name as an individual that they had

91 never met and to judge how well they thought the trait would describe that person. Participants saw each name once for a total of 72 trials. The 36 traits were presented twice, in a random order each time, once with a sonorant name and once with a voiceless stop name of the same gender

(the order of which was counterbalanced across participants). Pairing of traits with female vs. male names was counterbalanced across participants.

Results

We analyzed trials presenting traits from each of the six personality factors separately using mixed effects linear regressions. Our dependent measure was the rated fit between the presented trait and the name. Trials presenting a trait from the low end of a given factor were reverse coded. The type of name presented on a trial (i.e., sonorant vs. voiceless stop) was included as a fixed effect using effects coding (sonorant names = 0.5; voiceless stop names = -

0.5). Results indicated that sonorant (voiceless stop) names were judged as being higher (lower) in Agreeableness [0.41, 1.07] and Openness [0.15, 0.74]. Note that this effect was nearly statistically credible for Conscientiousness [0.00, 0.66]. Conversely, voiceless stop (sonorant) names were judged as being higher (lower) in Extraversion [-0.82, -0.15] (see Table 8).

Table 8

Resulting name type coefficients of linear regressions predicting fit ratings of sonorant (voiceless stop) names for traits from the high (low) end of each personality factor, in Experiment 2.

Factor B SE 95% CIs

Honesty-Humility 0.26 0.15 [-0.04, 0.56]

Emotionality 0.22 0.22 [-0.19, 0.64]

Extraversion -0.48 0.19 [-0.82, -0.15]*

Agreeableness 0.74 0.18 [0.41, 1.07]*

Conscientiousness 0.33 0.18 [0.00, 0.66]

Openness 0.44 0.15 [0.05, 0.74]*

* 95% Credible Interval does not include zero.

In a set of supplementary analyses, we added name gender (effects coded; female names

= 0.5; male names = -0.5), rated familiarity, and all interactions, to each of the models.

According to these models, sonorant (voiceless stop) names were still judged as being higher

(lower) in Agreeableness [0.41, 1.04] and Openness [0.11, 0.71]. Additionally, voiceless stop

(sonorant) names were still judged as being higher (lower) in Extraversion [-0.78, -0.11].

Participants also rated more familiar names as being higher in Honesty-Humility [0.01, 0.22] and

Extraversion [0.04, 0.25]. Finally, female (male) names were rated as being higher (lower) in

Emotionality [0.90, 1.54], Conscientiousness [0.44, 1.06] and Agreeableness [0.26, 0.93]. There were no statistically credible interactions.

Discussion

The results of Experiment 2 indicated that even when names were considered in isolation, sonorant and voiceless stop names had distinct associations with personality factors. In particular, sonorant names were judged as being higher in Agreeableness and Openness, while voiceless stop names were judged as being higher in Extraversion. In the next study we examine a potential implication of this effect: that these associations could actually emerge in the real world, in the personalities of individuals with names containing sonorants or voiceless stops.

Experiment 3

We next examined whether the effects observed in Experiments 1 and 2 might have real world effects on personality. Recent work has suggested that individuals might adjust their appearance to match stereotypes of their name (Zwebner, Sellier, Rosenfeld, Goldenberg, &

Mayo, 2017). Although it seems somewhat implausible, in order to understand the real-world limits of the effect we observed in the first two laboratory-based experiments reported here we felt we should test whether a similar process could take place with personality, with individuals subtly adjusting their personality to match sound symbolic associations of their names. To do so, we conducted a large-scale study examining relations between sound patterns in individuals’ real first names and their personalities, as measured by the HEXACO model of personality.

Method

Participants. There were 1,071 participants who took part online in exchange for financial compensation ($2 USD for approximately 15 minutes); these participants were recruited through Amazon Mechanical Turk. Participants were excluded if they failed any of our attention checks (19.61% of participants), did not provide their first name (7.75%), skipped more than five percent of items on either personality measure (7.00%) or did not provide their age

(1.87%). Our final sample included 843 participants (397 female, 443 male, 5 other; M Age =

36.63; SD = 11.47). We computed a power analysis using G* Power assuming a very small effect size (f2= 0.01) and α = .05. With seven predictors in the model, the power to detect an R2 increase caused by a single predictor was equal to 0.83.29

Materials and Procedure. Participants took part online through the website Qualtrics.

They completed an adjective-based and a statement-based personality inventory. In the adjective- based inventory, participants were asked to rate how well each of 60 traits applied to them on a five-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). These traits were the same as those used in our previous experiments, with an additional two from the high and the low ends of each factor (i.e., the two with the next highest loadings from Lee & Ashton, 2008, chosen with the same considerations as in Experiment 1). The only exception is that we bypassed several traits from the low-end of Honesty-Humility that related to Humility, in order to include traits related to Honesty. The statement-based inventory was the well-validated 100-item

HEXACO personality inventory revised (HEXACO-100; Lee & Ashton, 2016). It consists of

100 statements about the participant that are rated on a five-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). Of these, 16 pertain to each of the six personality factors, with an additional four pertaining to the interstitial scale of altruism (not analyzed here). Participants always completed the adjective-based inventory first, followed by the statement-based inventory.

The items for each inventory were presented in a random order. Note that self-report measures of

HEXACO personality have been shown to correspond with peer-reports (Lee & Ashton, 2006;

2016; 2017).

29 Note that this power analysis was computed as though the regressions to be run were univariate (rather than multivariate).

In addition to these personality inventories, participants were asked to provide their first name and a description of its pronunciation. They were also asked if they had an alternative name (e.g., shortened version of their first name or nickname; henceforth nickname), to describe its pronunciation, and to rate on a five-point scale how often they are addressed by that nickname. Finally, participants completed a demographic questionnaire including age, gender, ethnicity and education level.

Results

We first examined the validity of our personality measures. To begin, we computed correlations between participants’ scores on each of the six factors as measured by the adjective- or the statement-based inventory measure (see underlined cells in Table 9). Each of these were significant and of a moderate to high effect size. These correlations were also higher than any of the other 30 correlations between adjective- and statement-based scores. In addition, we examined if previously reported gender differences emerged in our data. Lee and Ashton (2004; see also Lee & Ashton, 2006) reported that women self-reported as being higher than men in

Honesty-Humility, Emotionality and Conscientiousness; conversely, men self-reported as being higher than women in Openness to experience. We found that women in our sample indeed scored higher than men on Honesty-Humility [0.14, 0.27; 0.19, 0.32], Emotionality [0.28, 0.40;

0.36, 0.48], and Conscientiousness [0.01. 0.14; 0.01, 0.15], on both the adjective- and the statement-based measures, respectively. Though we did not find the opposite pattern for

Openness, note that this was the smallest effect reported by Lee and Ashton (2004). Thus, we were satisfied with the validity of our personality data. The data also showed acceptable to excellent reliability (see Table 10).

96 Table 9

Zero-order correlations among predictor and outcome variables in Experiment 3. Correlations between the adjective- and statement-based

measures of a personality factor are underlined.

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

1. Age ---

2. Gender .12** ---

3. Proportion Sonorant .02 .18** ---

4. Proportion Voiceless Stop .01 -.04 -.24** ---

5. Starts with Sonorant .02 .19** .59** -.08* ---

6. Starts with Voiceless Stop .02 .07 -.17** .64** -.29** ---

7. Adjective H .21** .21** .05 .09** .08* .07* ---

8. Adjective E -.00 .34** .09* .02 .11** .05 .09** ---

9. Adjective X .11** -.02 -.04 .03 .00 .01 .33** -.17** ---

10. Adjective A .13** .06 -.02 .10** .01 .10** .54** .09* .27** ---

11. Adjective C .10** .08* -.01 .04 -.01 .07 .50** -.14** .41** .31** ---

12. Adjective O .04 -.01 -.07* .04 -.03 .03 .19** -.05 .21** .10** .21** ---

13. State H .27** .26** .03 .05 .02 .09* .47** .13** -.00 .32** .10** .05 ---

14. State E -.02 .42** .12** -.02 .11** .04 .01** .71** -.16** .00 -.11** -.07 .06 ---

15. State X .11** -.07 -.06 .06 -.03 .03 .29** -.23** .88** .26** .40** .23** .01 -.24** ---

16. State A .08* -.04 -.02 .05 -.01 .02 .42** -.12** .33** .74** .24** .07* .35** -.21** .37** ---

17. State C .08* .08* -.03 .06 -.03 .06 .41** -.13** .33** .23** .81** .27** .23** -.09** .37** .22** ---

18. State O .10** .04 -.07* .03 -.02 .02 .14** .05 .22** .06 .12** .62** .11** -.03 .24** .20** .21** **p < .05; **p < .01; ***p < .001

Table 10

Mean, Standard Deviation (calculated across subjects) and reliability for each of the personality inventories used in Experiment 3.

Factor Mean SD Cronbach’s α

Adjective-Based Inventory

Honesty-Humility 4.12 0.54 .85

Emotionality 3.23 0.60 .77

Extraversion 3.27 0.87 .91

Agreeableness 3.65 0.57 .78

Conscientiousness 4.00 0.65 .89

Openness 3.70 0.53 .76

Statement-Based Inventory

Honesty-Humility 3.50 0.71 .88

Emotionality 3.20 0.65 .86

Extraversion 3.16 0.77 .91

Agreeableness 3.18 0.65 .89

Conscientiousness 3.77 0.58 .86

Openness 3.58 0.63 .85

We then created a phonetic transcription for each participant’s first name and nickname.

When available, the American English transcription from the Carnegie Mellon Pronouncing

Dictionary (Weide, 2005) was used, via the website Lingorado

(http://www.http://lingorado.com/ipa/). For names not available in that source, we used

98 participant reported pronunciation to make a best guess at the names’ transcriptions. For participants who rated the frequency by which they go by their nickname as a five out of five, their nickname was used in the analyses instead of their first name. Names were analyzed based on the kinds of consonants they contained; this was quantified in two ways. As in Sidhu and

Pexman (2015), we calculated the proportion of total consonant phonemes in each name that were sonorants or voiceless stops. In addition, we categorized names based on their initial consonant phoneme (ignoring any preceding vowels), into those beginning with a sonorant, a voiceless stop, or neither (see Slepian & Galinsky, 2016). This was done via two effects coded

(and shifted) variables (i.e., names that begin with a sonorant vs. all other names; names that begin with a voiceless stop vs. all other names).

We analyzed the data using six multiple multivariate regressions (i.e., one for each personality factor). The two outcome variables in a given model were a participant’s scores on the adjective- and statement-based measures of a single personality factor. This was calculated by taking a mean of ratings on each factor, after reverse coding the appropriate items. The advantage of this approach (i.e., one multivariate regression vs. two univariate regressions for a given factor) is that the covariance in errors between two measures of the same personality factor is taken into consideration. All models also included participants’ age and gender. Gender was effects coded (and shifted; female = positive; male = negative). Note that we restricted analyses to individuals who identified as either male or female, as the frequency of the “other” category was not large enough to be analyzed (with only five observations), and we did not wish to force these participants into the male or female categories. Our predictors of interest were our four measures of name phonology, which were included in each model. For zero-order correlations among predictors, see Table 9. The results of these regressions can be found in Table 11. Results

99 indicated only one statistically credible relationship of name phonology on personality: individuals with a higher proportion of voiceless stop consonants in their names tended to have higher scores on the adjective-based measure of Honesty-Humility [0.01, 0.19].

100

Table 11

Resulting coefficients of multivariate linear regressions predicting participant scores for each personality factor as a function of name phonology, in Experiment 3.

Factor B SE 95% CIs B SE 95% CIs

Proportion Sonorant Proportion Voiceless Stops

Honesty-Humility

Trait Measure 0.01 0.04 [-0.07, 0.10] 0.10 0.05 [0.01, 0.19]*

Statement Measure 0.01 0.04 [-0.07, 0.09] 0.03 0.05 [-0.06, 0.12]

Emotionality

Trait Measure 0.00 0.04 [-0.08, 0.09] 0.02 0.05 [-0.07, 0.10]

Statement Measure 0.03 0.04 [-0.05, 0.11] -0.01 0.04 [-0.10, 0.07]

Extraversion

Trait Measure -0.06 0.05 [-0.15, 0.02] 0.01 0.05 [-0.09, 0.11]

Statement Measure -0.05 0.04 [-0.13, 0.04] 0.04 0.05 [-0.05, 0.14]

Agreeableness

Trait Measure -0.03 0.04 [-0.12, 0.05] 0.06 0.05 [-0.03, 0.16]

Statement Measure 0.01 0.04 [-0.08, 0.09] 0.06 0.05 [-0.03, 0.16]

Conscientiousness

Trait Measure -0.02 0.04 [-0.11, 0.07] -0.00 0.05 [-0.09, 0.10]

Statement Measure -0.02 0.05 [-0.11, 0.07] 0.03 0.05 [-0.06, 0.13]

Openness

Trait Measure -0.08 0.05 [-0.17, 0.01] 0.02 0.05 [-0.08, 0.11]

Statement Measure -0.09 0.05 [-0.18, -0.00] 0.01 0.05 [-0.08, 0.10]

101

Begins with a Sonorant Begins with a Voiceless Stop

Honesty-Humility

Trait Measure 0.05 0.05 [-0.05, 0.15] 0.01 0.06 [-0.11, 0.13]

Statement Measure -0.02 0.05 [-0.11, 0.08] 0.06 0.06 [-0.06, 0.17]

Emotionality

Trait Measure 0.06 0.05 [-0.04, 0.16] 0.04 0.06 [-0.08, 0.16]

Statement Measure 0.03 0.05 [-0.07, 0.12] 0.04 0.06 [-0.07, 0.16]

Extraversion

Trait Measure 0.06 0.05 [-0.05, 0.16] 0.01 0.06 [-0.12, 0.13]

Statement Measure 0.02 0.05 [-0.08, 0.12] 0.01 0.06 [-0.12, 0.13]

Agreeableness

Trait Measure 0.05 0.05 [-0.05, 0.15] 0.07 0.06 [-0.05, 0.20]

Statement Measure -0.01 0.05 [-0.12, 0.09] -0.03 0.06 [-0.15, 0.10]

Conscientiousness

Trait Measure 0.01 0.05 [-0.10, 0.11] 0.08 0.06 [-0.05, 0.20]

Statement Measure -0.02 0.05 [-0.13, 0.08] 0.03 0.06 [-0.09, 0.15]

Openness

Trait Measure 0.02 0.05 [-0.08, 0.12] 0.01 0.06 [-0.12, 0.13]

Statement Measure 0.03 0.05 [-0.07, 0.14] 0.01 0.06 [-0.11, 0.13]

* 95% Credible Interval does not include zero.

In a supplementary set of analyses, we examined whether the phonemes present in nicknames predicted personalities. Of the 335 participants who supplied a nickname, we

102 eliminated two who did note rate the frequency with which they went by that name, and 13 who rated that frequency a one out of five. For the remaining data, analyses were conducted in the same manner as described for first names. Results indicated one statistically credible relationship between nickname phonology and Honesty-Humility: individuals with a higher proportion of voiceless stop consonants in their names tended to have higher scores on the adjective-based measure of Honesty-Humility [0.04, 0.33]. There were also several statistically credible relationships between nickname phonology and Openness: individuals with a higher proportion of voiceless stop consonants in their nicknames tended to have higher scores on the adjective- based [0.09, 0.40] and statement-based measures of Openness [0.01, 0.32]; also, individuals with nicknames that started with a sonorant consonant tended to have lower scores on the statement- based measure of Openness [-0.41, -.0.04], see Table 12.

103

Table 12

Resulting coefficients of multivariate linear regressions predicting participant scores for each personality factor as a function of nickname phonology, in Experiment 3.

Factor B SE 95% CIs B SE 95% CIs

Proportion Sonorant Proportion Voiceless Stops

Honesty-Humility

Trait Measure 0.05 0.08 [-0.10, 0.20] 0.19 0.08 [0.04, 0.33]*

Statement Measure 0.06 0.07 [-0.09, 0.20] 0.05 0.07 [-0.09, 0.20]

Emotionality

Trait Measure 0.01 0.08 [-0.13, 0.16] 0.06 0.08 [-0.09, 0.21]

Statement Measure 0.05 0.07 [-0.09, 0.19] 0.07 0.07 [-0.06, 0.21]

Extraversion

Trait Measure -0.13 0.08 [-0.28, 0.03] -0.06 0.08 [-0.22, 0.09]

Statement Measure -0.06 0.08 [-0.22, 0.10] -0.05 0.08 [-0.21, 0.11]

Agreeableness

Trait Measure -0.01 0.08 [-0.16, 0.14] 0.10 0.08 [-0.05, 0.25]

Statement Measure 0.06 0.08 [-0.09, 0.21] -0.05 0.08 [-0.20, 0.11]

Conscientiousness

Trait Measure 0.02 0.08 [-0.14, 0.18] 0.02 0.08 [-0.14, 0.17]

Statement Measure 0.07 0.08 [-0.09, 0.22] 0.06 0.08 [-0.10, 0.22]

Openness

Trait Measure -0.02 0.08 [-0.17, 0.14] 0.24 0.08 [0.09, 0.40]*

Statement Measure 0.07 0.08 [-0.09, 0.22] 0.17 0.08 [0.01, 0.32]*

104

Begins with a Sonorant Begins with a Voiceless Stop

Honesty-Humility

Trait Measure -0.05 0.09 [-0.23, 0.13] -0.18 0.10 [-0.38, 0.01]

Statement Measure -0.12 0.09 [-0.30, 0.05] -0.00 0.10 [-0.20, 0.19]

Emotionality

Trait Measure 0.02 0.09 [-0.16, 0.20] -0.02 0.10 [-0.22, 0.17]

Statement Measure -0.02 0.09 [-0.19, 0.14] -0.06 0.09 [-0.24, 0.13]

Extraversion

Trait Measure 0.13 0.10 [-0.06, 0.32] 0.06 0.11 [-0.15, 0.27]

Statement Measure 0.02 0.10 [-0.18, 0.21] 0.01 0.11 [-0.20, 0.23]

Agreeableness

Trait Measure 0.11 0.09 [-0.08, 0.29] 0.05 0.10 [-0.15, 0.26]

Statement Measure 0.04 0.10 [-0.15, 0.22] 0.08 0.11 [-0.13, 0.29]

Conscientiousness

Trait Measure -0.08 0.10 [-0.27, 0.11] 0.06 0.11 [-0.14, 0.28]

Statement Measure -0.11 0.10 [-0.30, 0.08] 0.05 0.11 [-0.16, 0.26]

Openness

Trait Measure -0.17 0.10 [-0.36, 0.02] -0.19 0.11 [-0.40, 0.02]

Statement Measure -0.23 0.10 [-0.42, - -0.14 0.11 [-0.35, 0.07]

0.04]*

* 95% Credible Interval does not include zero.

105

Discussion

The results of this study do not provide much evidence that the phoneme-personality associations of an individual’s name are related to their self-reported personality. We only observed one statistically credible relationship–between proportion of voiceless stops and adjective-based Honesty-Humility. Note that this was not an association that we observed in

Experiments 1 and 2. The lack of an association between names and self-reported personality is not altogether surprising. While research has shown that individuals might subtly change their appearance to match stereotypes of their names (Zwebner et al., 2017), someone’s personality is much less malleable (see Roberts & DelVecchio, 2000). And of course, when parents give a child a name, they do not yet have insight into their personality. However, there was evidence of a link between nicknames containing voiceless stops (sonorants) and high (low) Openness. Note that this too was not an association that emerged in Experiments 1 and 2. Nevertheless, it may be that because nicknames are generally given to individuals later in life, when their personality is more apparent, it is possible for that personality to influence nickname choice.

A potential implication of these results is that the associations observed in Experiments 1 and 2 do not derive from large scale patterns of real first name sound symbolism in the population. In the next experiment, we test another potential explanation for the results of

Experiments 1 and 2: that information associated with the names used as stimuli could have contributed to the effects we observed. We chose to use real names as stimuli in Experiments 1 and 2 in order to examine whether phonology would have an effect even when situated in the context of words with existing semantic information. This approach is contrary to many studies on sound symbolism that use invented nonwords in order to be able to isolate the effects of phonology. Thus, we felt that Experiments 1 and 2 offered a stringent test of name phonology–if

106 phonology had an effect even in the presence of associated information then it must be rather robust. Notably, the fact that we observed an effect of familiarity demonstrates that participants indeed accessed this existing information to some extent. However, the strength of this approach is also a potential downside. That is, using real names creates the possibility that the effects we observed were somehow driven by this existing information. It could be the case, for instance, that there are salient individuals in popular culture who have the names we used, and that this contributed to the observed associations. To examine this possibility, we next ran a version of the task using invented names. This served to isolate name phonology, and remove any possible impact of existing semantic or episodic knowledge of certain names.

Experiment 4

In Experiment 4 we once again tested for phoneme-personality sound symbolism in the lab, using the Likert scale rating task from Experiment 2. In this case, however, participants were presented with invented names (i.e., letter strings that could be pronounced but were not real names) instead of real names. This allowed us to test the extent to which phoneme-personality sound symbolic effects arise from phonology alone, without the influence of episodic knowledge.

Method

Participants. Participants were 60 undergraduate students (46 female; M age = 20.97; SD

= 4.29) at the University of Calgary who participated in exchange for course credit. As this study measured the same construct as Experiments 1 and 2, we chose to test the same number of participants as in those experiments. All participants reported English fluency and normal or corrected to normal vision.

Materials and Procedure. Each of the names used in Experiment 2 was transformed into an invented name in the following manner. Each sonorant (voiceless stop) in a given name was

107 replaced with another sonorant (voiceless stop). When possible, this was done by rearranging the consonants in a name (e.g., Abel to Aleb; /eɪbəl/ to /eɪləb/). It was occasionally necessary to replace the vowels in a given name to avoid creating a real name or word. Replacement vowels were always sound symbolically congruent with the consonants of a given name (i.e., back vowels with sonorants; front vowels with voiceless stops; D’Onofrio, 2013). As in Experiments 1 and 2, the vowels of invented names were not perfectly congruent with their consonants.

Nevertheless, sonorant names contained a significantly higher percentage of back vowels (M =

48.15; SD = 32.80) than did voiceless stop names (M = 1.39; SD = 8.33), t(70) = 8.29, p < .001.

In addition, sonorant names contained a significantly lower percentage of front vowels (M =

23.15; SD = 27.68) than did voiceless stop names (M = 71.76; SD = 32.08), t(70) = 6.88, p <

.001. See Table B2 in Appendix B for a list of invented names. A trained linguist ensured that all invented names were phonotactically legal in English.

The procedure was identical to that described for Experiment 2, except that here visual presentation of the invented names was accompanied by an audio recording of their pronunciation (to ensure they were processed with the intended phonology). These recordings were created using Apple’s text to speech software. Participants were told that the recordings they would hear would be invented names. In addition, because the stimuli were invented names, the familiarity task was removed. Participants were presented with each invented name once for a total of 72 trials. The 36 traits were presented twice, in a random order each time, once with a sonorant name and once with a voiceless stop name (the order of which was counterbalanced across participants). Finally, we added a gender assignment task in which participants indicated whether they thought each invented name was more likely to be a male or female name via button press.

108

Results

Analyses were conducted in the same manner as Experiment 2. Results indicated that sonorant (voiceless stop) names were judged to be higher (lower) in Honesty-Humility [0.07,

0.65], Emotionality [0.27, 1.05], Agreeableness [0.42, 1.14], and Conscientiousness [0.07, 0.76].

Conversely, voiceless (sonorant) stop names were judged as being higher (lower) in Extraversion

[-0.66, -0.04] (see Table 13).

Table 13

Resulting name type coefficients of linear regressions predicting fit ratings of sonorant (voiceless stop) invented names for traits from the high (low) end of each personality factor, in Experiment

Factor B SE 95% CIs

Honesty-Humility 0.36 0.14 [0.07, 0.65]*

Emotionality 0.66 0.20 [0.27, 1.05]*

Extraversion -0.33 0.15 [-0.66, -0.04]*

Agreeableness 0.76 0.19 [0.42, 1.14]*

Conscientiousness 0.43 0.18 [0.07, 0.76]*

Openness 0.00 0.15 [-0.30, 0.30]

* 95% Credible Interval does not include zero.

In a set of supplementary analyses we also included the gender assigned to a particular name by each participant (which was effects coded and shifted to a mean of zero within trials of each personality factor; female = positive; male = negative), as well as its interaction with name

109 type. According to these models, sonorant (voiceless stop) names were still judged to be higher

(lower) in Honesty-Humility [0.01, 0.17], Emotionality [0.09, 0.26], Agreeableness [0.12, 0.28] and Conscientiousness [0.02, 0.20]; voiceless stop (sonorant) names were still judged to be higher (lower) in Extraversion [-0.18, -0.01]. In addition, names that participants thought were female (male) were judged to be higher (lower) in Emotionality [0.10, 0.29], Agreeableness

[0.10, 0.29] and Conscientiousness [0.06, 0.23]. There were no statistically credible interactions between name type and assumed gender.

Finally, we examined whether the relationship between name type and each personality factor was mediated by the perceived gender of each name. This was done at the item level, using the proportion of participants who identified a name as being male as a measure of perceived gender. We used the “mediation” package in R to perform the analyses. Name type was the predictor variable; the perceived gender of each name was the mediator variable. The rated agreement between each name and traits from a given factor was the dependent variable. The analysis used a quasi-Bayesian Monte Carlo method with 10,000 samples to estimate the indirect path (i.e., from name type to perceived gender to personality factor) and the direct path (i.e., from name type to personality factor) for each factor separately. The results are shown in Table 14. All previous statistically credible results were found to have a significant direct path.

110

Table 14

Results of mediational analysis in Experiment 4. Analysis was performed at the item level, with name type as the independent variable, and perceived name gender as a mediator variable.

Personality Factor Indirect Path Direct Path

Honesty-Humility 0.01 [-0.05, 0.08] 0.28 [0.07, 0.50]*

Emotionality 0.02 [-0.10, 0.15] 0.38 [0.21, 0.55]*

Extraversion 0.01 [-0.04, 0.06] -0.27 [-0.49, -0.05]*

Agreeableness 0.02 [-0.10. 0.15] 0.41 [0.24, 0.59]*

Conscientiousness 0.02 [-0.09, 0.14] 0.25 [0.05, 0.45]*

Openness 0.01 [-0.04, 0.06] 0.00 [-0.23, 0.23]

* 95% Confidence Interval does not include zero.

Discussion

Even after eliminating any influence of existing information associated with names, we observed associations between sonorant names and high Honesty-Humility, Emotionality,

Agreeableness and Conscientiousness; and between voiceless stop names and high Extraversion.

This suggests the that the phoneme-personality effects we observed are due to qualities of the names’ phonemes, rather than real world associations of the names. In fact, we observed a greater number of associations for invented names in Experiment 4 than we did for real names in

Experiment 2. On the one hand, it is possible that the addition of auditory information served to enhance the effects. On the other hand, the greater number of sound symbolic effects with invented names may support the interpretation that nonwords are read in a different manner that emphasizes effects of phonology, and/or that the presence of existing information with regards to

111 real names attenuates effects of sound symbolism. Notably, mediational analyses found no evidence of these results being mediated by the perceived gender of the names. This helps narrow in on sound symbolism as a primary cause of these effects, rather than associations between name phonology and gender. We next explored another potential mediating factor in these effects: differences in the likability of sonorant vs. voiceless stop names.

Experiment 5

We next collected ratings of the likability of each name and invented name used in lab, in

Experiments 1, 2 and 4, and examined the viability of likability as a mediating factor for the observed phoneme-personality associations. One might propose that some of the effects observed thus far (in particular sonorant names being more Agreeable and Conscientious) could reflect an overall bias to simply prefer sonorant names and associate them with positive traits. That is, that the phoneme-personality associations observed in Experiments 1, 2 and 4 could simply be explained as a valence effect, without requiring the nuance of invoking specific personality factors.

Method

Participants. Based on previous lexical-semantic ratings tasks (e.g., Brysbaert et al.,

2014; Warriner, Kuperman, & Brysbaert, 2013) we set a target sample size of 20 ratings per name. Our sample of convenience consisted of 27 undergraduate students at the University of

Calgary who participated in exchange for course credit. All participants reported English fluency and normal or corrected to normal vision. We excluded four participants who reported not being able to hear the audio files on a post-experiment questionnaire. This left a total of 23 participants whose data were analyzed (20 female, one non-binary; M age = 19.35; SD = 1.58).

112

Materials and Procedure. The ratings were collected using the online survey platform

Qualtrics. Participants were shown all 72 invented names from Experiment 4, along with the audio file of their pronunciation, in a random order. Participants were instructed to think of the name as belonging to an individual that they had never met and to judge how likable they think that person would be. Each name appeared along with a rating scale below it from 1 (“very unlikable”) to 7 (“very likable”). Participants then performed the same task for all of the names presented in Experiment 2, again in a random order. These were only presented visually.

Results

We analyzed likability ratings for names and invented names in separate mixed effects linear regressions. Our dependent variable was likability rating. The type of name presented was effects coded (sonorant = .5; voiceless stop = -.5), as was the gender of the name (shifted in the case of invented names, female names = positive; male names = negative; not shifted in the case of real names, female names = .5; male names = -.5). Invented name gender was based on the average number of participants in Experiment 4 who classified a name as male or female, using

50% as a cutoff. We also included an interaction between these predictors. In the case of invented names, we only observed an effect of gender such that female names were rated as more likable than male names [0.07, 0.33]. The results for real names indicated that round names were rated as more likable than sharp names [0.01, 0.20], with an interaction suggesting that this was greater

[0.01, 0.32] for female names.

We next examined whether the relationship between name type and each personality factor was mediated by the rated likeability of each name. This was done at the item level, in the same manner as the mediational analysis in Experiment 4. The results are shown in Table 15. All statistically credible effects from Experiments 2 and 4 were found to have a significant direct

113 path and, with one exception, a non-significant indirect path. In Experiment 2 we discovered a significant indirect path for Extraversion, but this was in the opposite direction as the direct path

(i.e., inconsistent mediation; MacKinnon, Fairchild, & Fritz, 2007). That is, sonorant names were rated as being higher in likeability, and more likeable names were rated as being higher in

Extraversion. However voiceless stop names were judged as being higher in Extraversion.

114

Table 15

Results of mediational analyses in Experiments 2 and 4. Analyses were performed at the item level, with name type as the independent variable, and name likeability as a mediator variable.

Personality Factor Indirect Path Direct Path

Experiment 2

Honesty-Humility 0.04 [-0.04, 0.13] 0.12 [-0.13, 0.36]

Emotionality 0.04 [-0.04, 0.13] 0.08 [-0.17 0.32]

Extraversion 0.13 [ 0.03, 0.25]* -0.47 [-0.68, -0.26]*

Agreeableness 0.01 [-0.07, 0.09] 0.44 [0.23, 0.67]*

Conscientiousness 0.03 [-0.05, 0.12] 0.22 [-0.01, 0.46]

Openness -0.04 [-0.13, 0.03] 0.36 [0.13, 0.60]*

Experiment 4

Honesty-Humility 0.00 [-.04, 0.05] 0.29 [0.08, 0.51]*

Emotionality 0.02 [-0.02, 0.09] 0.38[0.17, 0.60]*

Extraversion 0.04 [-0.03, 0.13] -0.30 [-0.52, -0.08]*

Agreeableness 0.04 [-0.03. 0.12] 0.40 [0.20, 0.60]*

Conscientiousness 0.05 [-0.04, 0.15] 0.22 [0.02, 0.43]*

Openness -0.01 [-0.06, 0.03] 0.02 [-0.22, 0.25]

* 95% Confidence Interval does not include zero.

Discussion

We found no effect of name type on likability for invented names but found that real sonorant names were rated as being more likeable than real voiceless stop names. This effect was

115 greater for female names. However, we found no evidence of likability mediating the effects observed in Experiments 1, 2 and 4. Thus, the relationships observed between phonemes in names and personality are not reducible to a valence effect. Notably, we discovered a statistically significant direct path for each of the significant name-personality sound symbolism findings in

Experiments 2 and 4 using a different statistical approach (i.e., item level mediational analysis vs. trial level mixed effects Bayesian regression) demonstrating that those effects are robust across different analytical techniques.

General Discussion

The purpose of the present study was to examine if there are sound symbolic associations between certain phonemes and personality factors. In Experiments 1, 2 and 4 we investigated whether first names containing sonorants vs. voiceless stops were differentially associated with personality factors from the HEXACO model. We did this by asking participants to choose between a pair of names as being most likely to possess a certain trait (Experiment 1), to rate the fit between an individual name and trait (Experiment 2) and to rate the fit between an individual invented name and trait (Experiment 4). Across these diverse procedures, we found evidence that sonorants and voiceless stops were indeed associated with different personality factors (Figure 7).

The most robust associations were between sonorant names and high Agreeableness

(Experiments 1, 2 and 4), Emotionality (Experiments 1 and 4) and Conscientiousness

(Experiments 1 and 4). We also found a robust association between voiceless stop names and high Extraversion (Experiments 2 and 4). While previous papers have investigated associations with isolated personality traits, the work here represents a comprehensive description of the relationships between sonorants/voiceless stops and broader personality space.

116

Figure 7. The results of Experiments 1, 2 and 4, for each personality factor. The x-axis corresponds to the 95% CIs for intercepts of each model in Experiment 1 and the coefficients of the name type predictors in Experiments 2 and 4. Positive values indicate that sonorants

(voiceless stops) were associated with traits from the high (low) end of a factor; negative values indicate the opposite.

Although the patterns of results we observed were generally consistent across experiments, there were some differences between experiments and this may simply be due to random variation. Alternatively, it could be that there is something systematic at play. For instance, the lone association between voiceless stops and high levels of a factor (i.e.,

117

Extraversion) only emerged when single names (rather than pairs) were presented (i.e., in

Experiments 2 and 4, but not in 1). While there may be something informative to such a pattern, we are wary of over interpretation.

It is also important to note that vowels may have contributed to the effects we observed, as sonorant names tended to have a higher percentage of back vowels, while voiceless stop names tended to have a higher percentage of front vowels. Thus the types of consonants and vowels in each name were not independent and we cannot make inferences about their individual contributions in isolation. It is noteworthy, however, that previous work on the maluma/takete effect has suggested that consonants are more influential than vowels (Nielsen & Rendall, 2011;

Ozturk, Krehm, & Vouloumanos, 2013). Nevertheless, while the focus here was on sonorant and voiceless stop consonants, future research might examine the effects of a variety of phonemes, potentially taking an unconstrained big data approach (as in Westbury et al., 2018).

A notable result is that the phoneme-personality associations emerged for real word stimuli in Experiments 1 and 2. This provides evidence that sound symbolism effects can emerge for real words with existing associated information. While that has been replicated in the case of name-shape associations (Sidhu & Pexman, 2015; Sidhu, Pexman, Saint-Aubin, 2016), it is less well established for name-personality associations. This latter type of association is particularly noteworthy, because while existing name information may not interfere with a round/sharp shape decision, it would be much more relevant to a decision about personality (i.e., recalling your friend Molly likely has little impact on a decision between a rounded and a jagged silhouette, but could be much more influential on a decision regarding Agreeableness). Importantly, these results are not restricted to real names, as we also observed phoneme-personality associations with invented names in Experiment 4. This supports the associations being due to the phonemes’

118 qualities, rather than real world associations of the names. Indeed, the fact that more associations were observed with invented names in Experiment 4 than with real names in Experiments 1 and 2 supports the notion that existing information can attenuate effects of sound symbolism.

Nevertheless, the finding that sound symbolism can affect the processing of real words has broad implications for language processing, as sound symbolically associated perceptual and/or semantic features could affect this process. Of course, future research will be needed to explore this beyond name stimuli.

In Experiment 3, we examined whether these associations emerged in a large scale analysis of self-reported personality in the real world. That is, we examined whether the proportion and/or initial appearance of sonorants and/or voiceless stops in a person’s name predicted their self-reported scores on the HEXACO factors, testing for sound symbolic associations like those observed in the lab. We did not find any evidence of this. In fact, we found evidence of only one association: individuals with a higher proportion of voiceless stops scored slightly higher on the adjective-based measure of Honesty-Humility. We are inclined to interpret this as a Type 1 Error, at least until the association is replicated. Interestingly, we found several pieces of evidence to suggest that there may be an association between voiceless stops

(sonorants) in nicknames and high (low) Openness. Note that this is not an association we observed in lab.30 Nevertheless, nicknames provide a more plausible instance in which to find a

30 Indeed this runs counter to what was observed in Experiment 2. We had proposed that individuals might subtly alter their personality to match the stereotypes associated with their names. However, there could be other processes by which the phoneme-personality associations of a person’s name affect their personality. Some of these could result in a personality that is opposite to the associations of the phonemes in one’s name. For instance, having a name that is associated with low Openness could lead individuals to behave in a way that runs counter to those expectations. It is also important to note that the in-lab experiments used a restricted set of names that were controlled for length and frequency, while Experiment 3 sampled from an unconstrained set of names. This might also lead to a discrepancy in findings. Of course, this is purely speculative, and we believe that the effects observed in Experiment 3 should be replicated before they are interpreted further.

119 relationship between names and personality, because they are often given later in life (and/or must endure to later in life) and thus can be influenced by an individual’s personality. Future research should explore this in a larger sample. It would also be important for research to distinguish between different types of nicknames (e.g., shorter versions of a first name, kinship terms). We did not have the necessary information to do so here.

Mechanisms for Phoneme-Personality Sound Symbolism

The present results suggest that phonemes have sound symbolic associations beyond the perceptual dimensions typically studied (e.g., size and shape), extending to the relatively more abstract dimensions of personality. It is notable that personality was studied here at the factor level. That is, phonemes were shown to go along with traits that loaded onto the same higher order abstract factor, suggesting an association with that higher order property. This fits with recent work showing a sound symbolic association between vowels and the abstract dimensions of social dominance (Auracher, 2017) and emotion (Rummer et al., 2014). It is also consistent with work showing that some perceptual sound symbolic associations may be driven by higher order abstract properties (Tzeng, Nygaard, & Namy, 2016). Demonstrating sound symbolism for abstract dimensions is important because it greatly broadens the purview of sound symbolism.

While the same core set of perceptual dimensions (e.g., shape and size) have been well studied as associated dimensions in sound symbolism, there are many potential abstract dimensions that have yet to be explored.

Abstract sound symbolism provides something of a challenge to explain, given phonemes’ lack of perceptual features. Exploring the possible mechanisms by which phonemes could be associated with abstract dimensions is informative as it broadens the scope of mechanisms that could underlie sound symbolism. Sidhu and Pexman (2018) laid out several

120 potential mechanisms for sound symbolism (see Table 16) and we will discuss here how two of these in particular (i.e., statistical co-occurrence and shared properties) could account for the effects we observed.

Table 16

Potential mechanisms underlying sound symbolic associations from Sidhu and Pexman (2018).

Mechanism and Description

Statistical Co-occurrence. Phonemes (or component features of phonemes; e.g., high pitch, a

component feature of high-front vowels) co-occurring with certain kinds of stimuli in the

world might lead to an internalization of those patterns and thus an association. For instance,

smaller things tend to resonate at a higher frequency, potentially explaining the association

between high-front vowels (which have a higher fundamental frequency) and small objects

(see Spence, 2011).

Shared Properties. Phonemes and associated stimuli could share some property in common,

be it perceptual or conceptual. Note that this shared property might require some element of

metaphor or analogy in order for it to apply across different modalities. For instance, the

abrupt onset of airflow when articulating a voiceless stop might lead to associations with the

abrupt changes in direction in the outline of a sharp shape (see Shinohara & Kawahara, 2012).

Neural Factors. The brain may be structured in such a way as to create an association

between stimuli from different modalities. For instance, there may be a neural link between

hand grasp posture and articulatory muscles. This might lead to an association between objects

inviting smaller grips, and the smaller articulations of high-front vowels (see Vainio,

Schulman, Tiippana, & Martti Vainio, 2013).

121

Evolutionary Factors. Evolution may have lead to organisms developing associations

between certain kinds of stimuli, if those associations provided a survival advantage. For

instance, organisms might be predisposed to associate higher pitches with the smaller

organisms producing those pitches (Ohala, 1994).

Language Patterns. Words for certain types of properties may disproportionately contain

some types of phonemes more than others. Individuals might then internalize that pattern and

come to associate these phonemes with such a property. For example, the onset gl- (i.e., a

phonestheme; see Bergen, 2004) appears in many words related to light, and individuals make

use of this word when asked to create words related to that property (Magnus, 2000).

One explanation that could be generated via the statistical co-occurrence account is that one tends to, for instance, encounter individuals with a sonorant name having a highly Agreeable personality. Over time repeated exposure to this co-occurrence might be internalized and lead to a phoneme-personality association. However, the results from the large scale personality analysis in Experiment 3 suggest that such a pattern does not exist in the population at large. Thus, we do not believe that phoneme-personality associations derive from such a large scale co-occurrence.

One might speculate that co-occurrences could exist in names besides real first names, especially when these names are given to individuals in light of their personality. For example, if agreeable fictional characters tended to be given sonorant names, this could represent a statistical co- occurrence. One might even speculate that a small number of extremely well-known fictional exemplars (or perhaps cultural icons) might go a long way to creating an association. A co- occurrence could also arise through nicknames if, for instance, sonorant nicknames tended to be given to those with agreeable personalities. While we found no evidence of nickname phonology

122 reflecting any of the patterns we observed in Experiments 1, 2 or 4, that analysis was low powered due to only a subset of our respondents providing a nickname. If such co-occurrences did exist in nicknames, they could contribute to phoneme-personality associations. Nevertheless, such a mechanism assumes an association exists to begin with (i.e., in order for many individuals to be given congruent nicknames, an association must already exist to drive such a pattern), and thus could not explain the origin of the associations observed here. However fictional or nickname patterns could be the final link in a causal chain that ultimately creates the association in an individual, or perhaps serve to “signal boost” the association. This highlights the fact that multiple mechanisms could play a role in creating these associations, at different points in causation.

There may be other manners in which phoneme features and personality factors could co- occur (i.e., outside of name stimuli) that could explain the origin of these effects. One possibility could be the tendency to use particular tones in certain emotional contexts. For instance, adults in distress tend to use harsh and punctuated voicings (Rendall, 2003). This tendency might lead to an association between soft, non-punctuated phonemes and traits related to the opposite of distress (e.g., perhaps those of high Agreeableness). Another possibility might be that the phoneme-personality associations observed in Experiments 1, 2 and 4 are mediated by the statistical co-occurrence of certain phonemes and gender. That is, certain phonemes being more common in female or male names (see Cassidy, Kelly, & Sharoni, 1990), could lead to those phonemes becoming associated with stereotypically female or male personalities, respectively.

However, note that the perceived gender of invented names did not mediate the effect of name type in Experiment 4.

123

Besides statistical co-occurrence accounts, it could be that phonemes and personality factors share some property in common. Because phonemes and personality are of a fundamentally different nature (i.e., acoustic/articulatory stimuli and abstract constructs), this would likely involve some amount of metaphor. Kawahara et al. (2015) speculated that the link between sonorants/obstruents and approachable/unapproachable personalities had to do with the acoustic properties of the phonemes relating to either personality metaphorically. That is, the abruptness of obstruents could metaphorically capture an unapproachable personality while the more smooth sonorants could capture an approachable personality. Something similar could be invoked to explain the associations here. For instance, acoustic smoothness could metaphorically map onto smooth social interactions with highly Agreeable individuals. Quick changes in acoustics (i.e., for voiceless stops) could metaphorically map onto energetic properties of highly

Extraverted individuals. Of course, these suggestions are highly speculative, and would require future research. Nonetheless, some evidence for this interpretation comes from studies on the connotative-semantic properties evoked by different types of phonemes. For instance, the sounds of sonorants evoke connotations of mellowness, passivity and delicacy; while voiceless stops evoke connotations of harshness, activity and ruggedness (Bozzi & Flores D’Arcais, 1963;

Greenberg & Jenkins, 1966). These may serve to connect sonorants with high Agreeableness and

Emotionality; and voiceless stops with high Extraversion and low Emotionality. Note that many of the mechanisms described so far can be invoked to explain an association between phonemes and some, but not all, of the associated personality factors. This could suggest that there are multiple mechanisms at play in the associations we observed.

As another possibility, it may be that phoneme-personality effects are determined at a higher level of abstraction (see Tzeng et al., 2016). That is, that phonemes and personality factors

124 could both be associated with some higher order property. While our results here suggest that this higher order factor is not likability, future research might explore other possibilities such as the two higher order properties suggested by Osgood, Suci and Tannenbaum (1963) in addition to valence: potency (i.e., strong-weak) and activity (i.e., active-passive).

Lastly, future research might examine whether the phoneme-personality associations found here actually mediate perceptual sound symbolism effects. Recent work has suggested that some perceptual sound symbolism associations are mediated by shared higher order semantic properties (Tzeng et al., 2016). It may be that the personality factors studied here are those shared properties, or at least are similar in meaning to those properties. While purely speculative, it is interesting to note that round shapes like those used in maluma/takete experiments are rated as being peaceful, tender, relaxed and friendly; while the sharp are rated as being aggressive, unfriendly and tough (Lindauer, 1990). This overlaps with some of the personality factors associated with sonorants and voiceless stops.

Real World Implications

An association between phonemes and the abstract dimension of personality has implications for the study of iconicity (i.e., words whose forms map onto their meanings). Much of this work has explored cases in which the form of a word resembles some perceptual property

(cf. Akita, 2011). For instance, ideophones (see Dingemanse, 2018) whose forms can convey sensory meanings via sound symbolism (e.g., the Japanese words goro and koro meaning a heavy and a light object rolling, making use of phoneme-weight sound symbolism). However, the results we observed here suggest that iconicity can also exist for more abstract meanings. Further, iconicity has been shown to benefit word learning (see Imai & Kita, 2014). Given the difficulty children have learning abstract language (see Ponari, Norbury, & Vigliocco, 2016), the results we

125 observed present a potential avenue for future research with regards to the acquisition of abstract language. That is, future work might look into how iconicity could be used to ground certain abstract meanings, and thus bootstrap their acquisition (see Imai & Kita, 2014; Perniss, Lu,

Morgan, & Vigliocco, 2018).

In addition, previous work has shown that impressions can be influenced by various features of a name (e.g., length, conventionality). The present studies suggest that another such feature is the sound symbolic associations of the phonemes in a name. Of course, we must remember that participants’ decisions in these experiments were made in the context of impoverished laboratory tasks. Indeed, participants had very little else on which to base their decisions, besides the phonology of the presented names. We expect that to the extent that individuals have additional information on which to base their judgments the effects of phonology on personality judgments would be attenuated. Nevertheless, there are everyday situations in which individuals are judged based on very little besides their name: for instance, in online communication. Future research should investigate the extent to which first impressions or resume evaluations can be influenced by the sound symbolism of a first name. Advertising could also make use of these associations when choosing names for products (see Klink & Athaide,

2012), or characters in advertisements.

Lastly, the associations observed in Experiments 1, 2 and 4 could have real world effects when individuals choose a name for targets with certain personalities. The most obvious instance of this would be when an author chooses a name for a character. They may–consciously or unconsciously–select a name that is congruent with the character’s personality, in order to highlight that personality for the audience (see Elsen, 2017; Kawahara, Noto, & Kumagi, 2017;

Smith, 2006). Future research might examine the extent to which the associations observed here

126 are present in works of fiction, as well as the impact of congruent/incongruent names on the reader’s experience of that fiction. It would also be interesting to explore the extent to which these associations affect differences in naming trends between boys and girls. For instance, research has shown that sonorants are more common in female vs. male names (Sidhu &

Pexman, 2015). This could be related to sonorants being associated with stereotypically female personality qualities such as high Agreeableness (see Huddy & Terkildsen, 1993).

Conclusion

We investigated whether phonemes have sound symbolic associations beyond the perceptual effects typically studied, using the more abstract dimensions that comprise the construct of personality. Across three laboratory studies we found sound symbolic associations for the phonemes in first names, with sonorants showing an association with high Emotionality,

Agreeableness, and Conscientiousness; and voiceless stops showing an association with high

Extraversion. These results suggest that any theory of sound symbolism should take into consideration associations between phonemes and more abstract dimensions.

127

Acknowledgements

This research was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) through a postgraduate scholarship to DMS and a Discovery Grant to PMP; and by Alberta Innovates: Health Solutions (AIHS) through a graduate scholarship to DMS. The authors thank Stephanie Archer for her help assessing the stimuli used in Experiment 4.

128

Chapter 4: Effects of Iconicity in Lexical Decision

Introduction

Traditional views of language have held that the relationship between the form of a word and its meaning is arbitrary (e.g., Hockett, 1963). That is, there is no special link between a word’s form (i.e., its orthography, phonology or articulation) and its meaning. However, arbitrariness may be only one possible kind of relationship between form and meaning (see

Dingemanse, Blasi, Lupyan, Christiansen, & Monaghan, 2015; Perniss & Vigliocco, 2014).

Language can also be non-arbitrary via iconicity: a resemblance between form and meaning, in which aspects of a word’s form map onto aspects of its meaning (e.g., Perniss, Thompson &

Vigliocco, 2010). In this paper we focus on phonological iconicity, wherein a word’s phonological form maps onto meaning.

Iconicity can occur in several ways. In onomatopoeia, the phonology of a word imitates the sound to which it refers. English examples include quack, boom and sizzle. It is also possible for a word’s form to map onto its meaning without direct imitation, but instead via association

(i.e., non-onomatopoeic iconicity). This is enabled by the phenomenon of sound symbolism, in which particular language sounds evoke associations with non-auditory properties (see

Lockwood & Dingemanse, 2015; Sidhu & Pexman, 2018a). For instance, individuals associate the vowels /ɪ/ and /ɑ/ with smallness and largeness, respectively (Newman, 1933; Sapir, 1929).

Thus, a word like shrimp could be considered iconic because its form evokes associations (i.e., smallness) that map onto aspects of its meaning.

It has been claimed that arbitrariness and iconicity both provide benefits to language.

Arbitrariness allows any form to refer to any meaning without requiring resemblance, which may not always be possible (e.g., for entirely abstract concepts; Dingemanse et al., 2015). Iconicity

129 facilitates language learning, in part by helping infants and toddlers associate speech sounds with their referents (see Imai & Kita, 2014, Laing, Viham, & Portnoy, 2017; Perniss, Lu, Morgan, &

Vigliocco, 2018; Perry, Perlman, Winter, Massaro, & Lupyan, 2017). In addition, it has been argued that the imitative, performative nature of iconic words makes language more direct and vivid (Dingemanse et al., 2015). In the present experiments we examined another potential advantage of iconicity: that iconic links between sound and meaning make these words easier to process.

Importantly, words are not entirely arbitrary or iconic. Rather, they fall on a spectrum from extremely arbitrary to extremely iconic–containing aspects of arbitrariness and iconicity

(see Dingemanse et al., 2015; Perniss & Vigliocco, 2014). For instance, while the word quack sounds like a duck quacking, there are many other ways to imitate this sound (e.g., mac in

Romanian, vak in Turkish, or prääks in Estonian), and thus the choice of quack is to some extent an arbitrary one. Indeed, other sets of phonemes might better imitate the sound of a duck. Also, while /ɪ/ in the word shrimp maps onto an aspect of its meaning, its other phonemes are seemingly arbitrary.31 A word that contained a greater proportion of small-associated phonemes might seem more iconic. Perry, Perlman, and Lupyan (2015; supplemented by Winter, Perlman,

Perry, & Lupyan, 2017) quantified the subjective iconicity of a large set of English words. They had participants rate a variety of English words in terms of their iconicity on a continuous scale and discovered that words existed along the entire spectrum.

There is a great deal of work to be done to understand the effects of such variation in iconicity on language processing. Triangle models of word recognition include three components:

31 We thank the editor for pointing out that there is a difference in the nature of the arbitrariness in quack and shrimp. There are only so many phoneme combinations that could imitate the sound of a duck, and so while there is some room for variation, the phonemes in quack are still motivated and not entirely arbitrary. Conversely, there is seemingly no restriction in the non-iconic phonemes in shrimp (i.e., /ʃ/).

130 a word’s meaning, its orthography and its phonology. In these models, a word’s meaning is accessed via two paths: directly from a word’s orthography, and indirectly from its orthography via its phonology (Harm & Seidenberg, 2004). While other models of word recognition exist

(e.g., dual route models; Coltheart et al., 2001), here we focus on triangle models because they specify a route by which meaning is retrieved via phonology. Note that links between each of the three components of triangle models are bidirectional. Additionally, the extent to which paths or components are emphasized varies depending on task context (Balota, Paul, & Spieler, 1999).

Evidence for this is the fact that phonological variables (e.g., spelling-sound regularity; Hino &

Lupker, 1996) play more or less of a role in word recognition when the orthographic- phonological path is prioritized to a greater or lesser extent.

Mappings between phonology and semantics are considered arbitrary in triangle models: they must be learned through experience.32 However, iconicity presents the possibility that some connections between phonology and semantics are not solely arbitrary. Such connections may emerge or exist naturally, or be easier to learn, by virtue of the inherent resemblance between phonology and semantics. Here we evaluate the possibility that iconic words possess extra and/or more direct links between phonology and semantics which will facilitate their processing. Note that this may be a quantitative and/or a qualitative difference in the nature of links for iconic as compared to non-iconic words.

Meteyard et al. (2015) examined the possibility that special links between phonology and semantics might make iconic words more resistant to aphasia. They compared processing of onomatopoeia and arbitrary words in aphasic patients using a variety of tasks. They found a

32 Though Harm and Seidenberg (2004) mention systematic non-arbitrary patterns in orthographic-semantic links (e.g., relations among words shared an affix) that could affect learning.

131 benefit for onomatopoeia on the tasks that prioritized the mapping of phonology onto semantics.

The authors also speculated that effects of iconicity may only be observable in individuals with developing or damaged language systems.

Peeters (2016) conducted an EEG study on the processing of Dutch onomatopoeia using an auditory lexical decision task. He found that onomatopoeia elicited a smaller N400, interpreted as reflecting facilitated lexical access for words with iconic mappings between form and meaning. However, despite this effect, there were no behavioural differences in the processing of onomatopoeic and arbitrary words. Peeters speculated that onomatopoeia may be processed both as linguistic stimuli and as environmental sounds (e.g., boom could be interpreted both as a word and the sound to which it refers). Peeters suggested that this could interfere with lexical decision where participants must judge stimuli to be linguistic stimuli or not.

Other evidence of iconicity potentially interfering with performance on a word recognition tasks comes from an analysis by Lupyan and Winter (2018). They analyzed data from a semantic decision task (abstract/concrete) collected by Pexman, Heard, Lloyd and Yap (2017), and found that higher iconicity actually led to a lower accuracy for more abstract items. Their interpretation was that words that are more iconic activate more specific semantic representations, which nudges participants towards a “concrete” (i.e., incorrect) decision.

In the present experiments we explored two questions. First, we examined whether the recognition of English iconic words is facilitated in a population of typical adults on a visual lexical decision task. Second, as a first step towards exploring the locus of such an effect, we varied the extent to which phonology was prioritized by the task.

In Experiment 1 we presented words varying in their iconicity in a visual lexical decision task (LDT: is this letter string a word?). While some previous studies have used auditory lexical

132 decision tasks (e.g., Meteyard et al., 2015), we elected for a visual lexical decision task because:

1) there is still evidence of phonological processing on such a task (e.g., Pexman, Lupker, &

Jared, 2001), and 2) using a more conservative test, and one that is more similar to everyday reading processes, allowed us to ensure that any effects would be more broadly applicable.

Stimuli in the LDT were presented clearly or visually degraded. Visual degradation prevents participants from making lexical decisions solely on the basis of orthographic information, and thus tends to increase the extent to which phonological information is recruited (Hino & Lupker,

1996). If the nature of the links between phonology and semantics for iconic words provide a benefit to word recognition, enhancing the extent to which phonology is recruited by the task may enhance this effect. In Experiment 2 we presented the same items in a phonological lexical decision task (PLDT: does this letter string sound like a word?). Phonology is explicitly emphasized in the PLDT, as participants are asked to make responses based on the phonology rather than the spelling of each item.

Experiment 1

Methods

Participants. Participants were 80 undergraduate students (61 female; M age = 21.01; SD

= 3.38) at the University of Calgary who participated in exchange for course credit. Participants reported English fluency and normal or corrected to normal vision.

Materials and Procedure. Stimuli were 120 real words and 120 nonwords. The real words were chosen from the iconicity ratings collected by Perry et al. (2015) and Winter et al.

(2017), in which words were rated on a scale from -5 (the word sounds like the opposite of its meaning) to 5 (the word was highly iconic), with 0 indicating that the word was arbitrary. In order to sample broadly from different types of words, we selected and matched words as per a

133 factorial design, though in the analyses we treated iconicity as a continuous variable. To that end, we selected words corresponding to three categories: non-iconic words, onomatopoeia, and non- onomatopoeic iconic words. Non-iconic words had iconicity ratings ≥ -0.50 and ≤ 0.50 (M =

0.12; Perry et al., 2015; Winter et al., 2017). Onomatopoeia had iconicity ratings ≥ 2.50 (M =

3.42) and had phonologies that imitated their meanings. Non-onomatopoeic iconic words had iconicity ratings ≥ 2.50 (M = 2.99) but had phonologies that we judged not to directly imitate their meanings (e.g., twist, fluff, slime). The three types of words were matched on length, log subtitle word frequency (Brysbaert & New, 2009), number of morphemes, orthographic

Levenshtein distance (OLD; Yarkoni, Balota, & Yap, 2008), phonological Levenshtein distance

(PLD; Yarkoni et al., 2008), mean bigram frequency, age of acquisition (Kuperman, Stadthagen-

Gonzalez, & Brysbaert, 2012), concreteness (Brysbaert, Warriner, & Kuperman, 2014), and initial syllable rime phonological consistency (Yap, 2007). In addition, each type contained 26 nouns and 14 verbs. See Table 17 for properties of each type. The 120 nonwords did not contain any pseudohomophones (i.e., nonwords that share phonology with a real word; e.g., brane) and were matched with the 120 real words on length, number of orthographic neighbours, mean bigram frequency and number of syllables. See Table C1 in Appendix C for a list of stimuli.

134

Table 17

Mean (SD) values of lexical and semantic variables for each of the word types presented in

Experiments 1 and 2.

Non-Iconic Onomatopoeia Non-Onomatopoeic Iconic

Iconicity 0.12 (0.27) 3.42 (0.50) 2.99 (0.40)

Length 4.88 (1.29) 4.90 (1.11) 4.90 (1.06)

Frequency 2.32 (0.80) 2.30 (0.50) 2.31 (0.73)

Number of Morphemes 1.03 (0.16) 1.00 (0.00) 1.03 (0.16)

OLD 1.72 (0.60) 1.72 (0.43) 1.73 (0.41)

PLD 1.51 (0.65) 1.49 (0.36) 1.52 (0.43)

Mean Bigram Frequency 1359.74 (636.88) 1340.30 (643.73) 1351.32 (655.26)

AoA 6.99 (2.67) 6.99 (1.77) 7.00 (1.89)

Concreteness 3.82 (1.15) 3.82 (0.43) 3.82 (0.60)

Phonological Consistency 0.87 (0.22) 0.87 (0.23) 0.88 (0.22)

Note. Frequency is log subtitle frequency; OLD is orthographic Levenshtein distance; PLD is phonological Levenshtein distance; AoA is average year at which word was acquired;

Phonological Consistency is the first syllable rime feedforward consistency.

Participants completed an LDT in which their task was to categorize a presented letter string as a word or a nonword. Each trial began with a fixation cross for 400 msec, followed by a blank screen for 200 msec, after which the target letter string was presented. Participants categorized stimuli as nonwords or words via keyboard press. Their response triggered a 550 msec blank screen, after which the next trial began. If participants made an error they saw the

135 word “Incorrect”, and heard a brief sound, during this blank screen. Participants wore sound attenuating headphones during the task. Stimuli were presented in a random order, in two blocks, with a break between blocks. Participants saw an equal proportion of each stimulus type in each block.

In addition, participants were randomly assigned to one of two presentation conditions: clear or degraded (40 participants in each). In the clear condition stimuli were presented normally. In the degraded condition we followed the approach taken by Yap, Lim and Pexman

(2015) to visually degrade the letter strings, rapidly alternating between the letter string and a mask of random symbols of the same length. This condition was run with a refresh rate of 144

Hz.

Results

Statistics. We used the packages "lme4" [version 1.1-18-1] (Bates, Maechler, Bolker, &

Walker, 2015), "afex" [0.23-0] (Singmann, Bolker, & Westfall, 2015), and "RePsychLing"

[0.0.4] (Baayen, Bates, Kliegl, & Vasishth, 2015) to perform our statistical analysis in R [3.5.1]

(R Core Team, 2016). We took a confirmatory approach and fit models including all fixed effects of interest. We developed each model’s random effects structure using the approach suggested by

Bates, Kliegl, Vasishth, and Baayen (2015). In brief, we began by fitting the model with all random slope terms for each fixed effect, and removed correlations among random effects if this did not converge. We then performed a principal components analysis on the random effects and simplified the structure based on the suggested number of components (Baayen et al., 2015). We also tested the inclusion of correlations among random effects, and the effects themselves, using likelihood ratio tests. The detailed procedure for model selection, along with code used for the entire process, can be found in the online supplementary materials: https://osf.io/ue7sv/. We

136 generated p-values using the package "lmerTest" [3.0.1] (Kuznetsova, Brockhoff, & Christensen,

2017). The "prediction" package [0.3.6] was used to generate marginal predictions. Throughout all results, we only report analyses based on real word trials.

We took the following approach to cleaning the data in these and all analyses of reaction time. First, we excluded all incorrect responses. Then trials with a reaction time less than 200 msec or greater than 3000 msec were removed. We then removed trials that were more than 2.5 standard deviations away from a participant’s mean. No more than 5.13% of trials were ever removed by this process. We took the same approach to cleaning the accuracy data except that incorrect responses were not excluded.

Reaction time. We ran a linear mixed effects model33 that predicted reaction time using condition (effects coded; -1 = clear presentation, +1 = degraded presentation), continuous iconicity (Perry et al., 2015; Winter et al., 2017) and their interaction. Length, frequency (log subtitle word frequency) and OLD were also included as control variables. This model revealed that there was no interaction between condition and iconicity (b = 1.82, p = .41). However, there was a main effect of iconicity such that reaction times were faster in response to words with higher iconicity (b = -13.65, p = .002). In particular, iconicity values of -1.5 and +1.5 SD resulted in predicted reaction times of 724.22 and 683.41 msecs, respectively. In addition, there was a main effect of condition, such that reaction times were faster in the clear condition (b = -45.77, p

= .02). In particular, the clear and degraded conditions resulted in predicted reaction times of

659.19 and 750.72, respectively. See Table 18 for a model summary.

33lmer(RT~Length+Frequency+OLD+Iconicity*Condition+(1|Subject)+(1|Word)) Note that all subsequent models use this same general structure with changes to the included fixed or random effects noted in text. The specific models used in each analysis can be found in the following OSF repository: https://osf.io/ue7sv/.

137

Table 18

Linear mixed effects regression model predicting LDT reaction time in Experiment 1.

Fixed Effect B S.E. sr2 t p

Intercept 709.87 19.39 36.62 < .001***

Control Variables

Length -4.60 6.59 .00 -0.70 .49

Frequency -36.83 4.56 .02 -8.07 < .001***

OLD 15.76 6.70 .003 2.35 .02*

Predictor Variables

Iconicity -13.65 4.25 .005 -3.22 .002**

Degradation Condition -45.77 19.04 .03 -2.40 .02*

Iconicity x Degradation Condition 1.82 2.20 .00 0.83 .41

Random Effect s2

Item Intercept 1573

Subject Intercept 28615

Notes. * p < .05, ** p < .01, *** p < .001. Marginal R2 = .05, computed using the Nakagawa and

Schielzeth method via the “r2glmm” package in R (Jaeger, 2017).

Accuracy. We ran a logistic mixed effects model that predicted accuracy and included the same fixed and random effects as in the main analysis of reaction time. This model revealed that there was no interaction between condition and iconicity (b = -0.01, p = .72). However, there was a main effect of iconicity such that correct responses were 1.25 times more likely for each 1 SD

138 increase in iconicity (b = 0.22, p = .005). There was not, however, a main effect of condition (b =

0.13, p = .12). See Table 19 for a model summary.

Table 19

Logistic mixed effects regression model predicting LDT accuracy in Experiment 1.

Fixed Effect B exp(B) S.E. sr2 Wald’s z p

Intercept 2.94 18.98 0.11 25.79 <.001***

Control Variables

Length 0.57 1.77 0.13 .004 4.56 <.001***

Frequency 0.79 2.21 0.09 .02 8.66 <.001***

OLD -0.24 0.79 0.12 .001 -1.95 .052

Predictor Variables

Iconicity 0.22 1.25 0.08 .002 2.80 .005**

Degradation Condition 0.13 1.14 0.09 .001 1.55 .12

Iconicity x Degradation Condition -0.01 0.99 0.04 .00 -0.36 .72

Random Effect s2

Item Intercept 0.52

Subject Intercept 0.44

Notes. * p < .05, ** p < .01, *** p < .001. Marginal R2 = .02, computed using the Nakagawa and

Schielzeth method via the “r2glmm” package in R (Jaeger, 2017).

139

Discussion

Participants were faster and more accurate when responding to words that were higher in iconicity. This suggests that iconicity does provide a benefit to word recognition. Interestingly, the effect of iconicity did not interact with condition, suggesting that any increased use of phonology due to visual degradation did not enhance the effect of iconicity. In the next experiment we explored the role of phonological encoding further, using a task that explicitly prioritizes phonology, by requiring participants to attend to the phonologies of each word (and not the spellings) in order to make a correct response.

Experiment 2

Methods

Participants. Participants were 48 undergraduate students at the University of Calgary who participated in exchange for course credit. Participants reported English fluency and normal or corrected to normal vision. We tested more participants than we intended to analyze (i.e., 40), because previous work with the PLDT has found that some participants do not comply with the instructions to emphasize phonology (e.g., Pexman, Lupker & Reggin, 2002). This noncompliance is evident in accuracy on pseudohomophone trials. Thus, we included the 40 participants (25 female; M age = 23.15; SD = 5.41) with the highest accuracy on pseudohomophone trials (i.e., correctly categorizing stimuli like brane as sounding like a real word; see below for details).34

Materials and Procedure. In addition to the stimuli from Experiment 1, we added 60 pseudohomophones and 60 additional nonwords. These pseudohomophones were created by altering the spelling of an existing word (e.g., cough to koff). As in Experiment 1, we selected and

34 Note that the to-be-reported significant effects remain significant when including all 48 participants.

140 matched words as per a factorial design, though we treated iconicity as a continuous variable in the analyses. To that end, pseudohomophones were created using the phonologies of 20 non- iconic, 20 onomatopeia, and 20 non-onomatopoeic iconic words. Stimuli requiring a “word” response and stimuli requiring a “nonword” response were matched on length, orthographic neighbourhood size, mean bigram frequency and number of syllables. See Table C2 in Appendix

C for a list of stimuli.

Participants completed a PLDT in which their task was to categorize a letter string as sounding like a word (i.e., having the phonology of a real word) or a nonword (i.e., not having the phonology of a real word). Except for the different decision criterion, trials were presented in the same manner as in the clear condition of Experiment 1.

Results

Data were analyzed using the same approach as in Experiment 1.

Reaction time. We ran a linear mixed effects model that predicted reaction time using iconicity as a predictor. The model also included length, frequency (log subtitle word frequency) and PLD (as this experiment prioritized phonology) as control variables. This model revealed an effect of iconicity such that reaction times were faster in response to words with higher iconicity

(b = -22.48, p < .001). In particular, iconicity values of -1.5 and +1.5 SD resulted in predicted reaction times of 780.33 and 712.90 msecs, respectively. See Table 20 for a model summary.

141

Table 20

Linear mixed effects regression model predicting PLDT reaction time in Experiment 2.

Fixed Effect B S.E. sr2 t p

Intercept 749.12 17.50 40.05 < .001***

Control Variables

Length -0.09 8.40 .00 0.23 .82

Frequency -52.78 6.34 .04 -8.19 < .001***

PLD 10.75 8.70 .001 1.00 .32

Predictor Variables

Iconicity -22.48 5.81 .008 -2.65 .009**

Random Effect s2

Item Intercept 2704

Subject Intercept 10894

Notes. * p < .05, ** p < .01, *** p < .001. Marginal R2 = .06, computed using the Nakagawa and

Schielzeth method via the “r2glmm” package in R (Jaeger, 2017).

We ran a supplementary analysis in which we combined the present results with those of the clear condition in Experiment 1 and tested for an interaction between task (LDT vs PLDT) and iconicity, to examine the impact of explicitly directing participants to focus on phonology.

The model also included length, frequency and OLD as control variables; it also included a random item slope for the effect of task. The interaction between task and iconicity was non- significant (b = -5.24, p = .06).

142

Finally, we ran a supplementary analysis on the pseudohomophone trials. We ran a model that predicted reaction time using the iconicity of the word on which each pseudohomophone was based. The model also included length, orthographic neighbourhood size, and mean bigram frequency of the pseudohomophones, and base word frequency (log subtitle frequency), as control variables. This model revealed an effect of base word iconicity, such that reaction times were faster in response to pseudohomophones that were based on words with higher iconicity (b

= -61.78, p = .04). In particular, base word iconicity values of -1.5 and +1.5 SD resulted in predicted reaction times of 1196.96 and 1011.61, respectively.

Accuracy. We ran a logistic mixed effects model that predicted accuracy using the same fixed and random effects as in the main analysis of reaction time. This model found no effect of iconicity on response accuracy (b = 0.18, p = .13), see Table 21.

143

Table 21

Logistic mixed effects regression model predicting PLDT accuracy in Experiment 2.

Fixed Effect B exp(B) S.E. sr2 Wald’s z p

Intercept 4.01 0.03 0.21 18.72 <.001***

Control Variables

Length 0.29 0.15 0.17 0.00 1.74 .08

Frequency 0.63 0.19 0.13 0.004 4.85 <.001***

PLD 0.08 0.36 0.17 0.00 0.43 .67

Predictor Variables

Iconicity 0.18 0.29 0.12 0.00 1.52 .13

Random Effect s2

Item Intercept 0.65

Subject Intercept 0.83

Notes. * p < .05, ** p < .01, *** p < .001. Marginal R2 = .005, computed using the Nakagawa and

Schielzeth method via the “r2glmm” package in R (Jaeger, 2017).

We ran a supplementary analysis in which we combined the present results with those of the clear condition in Experiment 1 and tested for an interaction between task and word type. The model also included length, frequency and OLD as control variables. This model found no interaction between task and iconicity (b = 0.00, p = .97).

Finally, we again ran a supplementary analysis on the pseudohomophone trials. We ran a model that predicted accuracy using the same fixed and random effects as in the analysis of pseudohomophone reaction time. This model revealed an effect of base word iconicity, such that

144 correct responses were 1.80 times more likely for each 1 SD increase in base word iconicity (b =

0.59, p = .004).

Discussion

In Experiment 2, we found that participants responded faster to words with higher iconicity. They did not, however, respond more accurately. Interestingly, an iconicity benefit also emerged in the processing of pseudohomophones. As the correct identification of these items would have relied on phonology, this speaks to the role of phonology in iconicity. We next examined responses to these words when presented in the broader context of a megastudy that included a wide variety of words.

English Lexicon Project Analysis

We performed an analysis examining average reaction time and accuracy in the English

Lexicon Project (ELP; Balota et al., 2007) LDT for the 120 words used in Experiments 1 and 2.

Results

We ran a linear model that predicted LDT reaction time in the ELP using iconicity as a predictor. The model also included length, frequency (log subtitle word frequency) and OLD as control variables. Note that this analysis was done at the item level, and therefore did not include random effects. This model found no effect of iconicity on reaction time (b = 11.00 p = .054).

The corresponding logistic model found no effect of iconicity on accuracy (b = 0.01 p = .46).

In a supplementary analysis, we explored the possibility that list context might have led to the difference in results between the ELP data and the non-degraded LDT from Experiment 1

(which both used the same task). While ELP participants were presented with a wide variety of words, two-thirds of the words in Experiment 1 were iconic. We examined whether this list context in the non-degraded LDT from Experiment 1 increased participants’ attention to iconicity

145 through a linear mixed effects model that included trial number and its interaction with iconicity, in addition to previously used control variables. We found a significant interaction between trial number and iconicity (b = -8.82, p = .002), such that iconicity had a greater effect on later trials

(see Figure 8). In particular, iconicity values of -1.5 and +1.5 SD resulted in predicted reaction times of 672.76 and 660.88 on trial 60 (of 240), and 688.20 and 629.91 on trial 180, respectively.

The corresponding logistic mixed effects model found no interaction between condition and trial number (b = 0.06, p = .34).

146

Figure 8. Plot showing the relationship between iconicity and predicted reaction time in the non- degraded LDT from Experiment 1. This was calculated separately for each quarter of the experiment (e.g., the first 60 trials that a participant saw represents the first quarter of the experiment). Each line shows the effect in a different quarter of the experiment.

Discussion

We found no effect of iconicity on LDT response times in the ELP. This contrasts with results from the non-degraded LDT in Experiment 1. One possibility is that list context plays a role. While two-thirds of the words in Experiment 1 were iconic, this was not true in the ELP.

Speaking to this is the fact that the effect of iconicity on reaction time in the non-degraded LDT

147 from Experiment 1 emerged over time, potentially as participants shifted their attention towards iconicity as a useful cue to a “word” response.

General Discussion

It had traditionally been assumed that the relationship between a word’s form and its meaning is arbitrary (e.g., Hockett, 1963), but a more recent perspective is that this relationship contains aspects of arbitrariness and iconicity (Dingemanse et al., 2015; Perniss et al., 2010).

Here we tested the possibility that recognition might be facilitated for relatively more iconic words. Indeed, we found that more iconic words were processed faster and more accurately in a visual LDT, and faster in a phonological LDT.

These experiments suggest that iconicity can confer an advantage in processing to typical individuals, in addition to people with aphasia (Meteyard et al., 2015). These results stand somewhat in contrast to a study on Dutch onomatopoeic words that found no effects in reaction time (Peeters, 2016). Note that stimuli in the Dutch study were presented auditorily, and that this difference in modality may interact with effects of iconicity (cf. Meteyard et al., 2015). In fact,

Peeters speculated that the lack of an iconicity effect may have been due to onomatopoeia being more difficult to identify as words (as opposed to environmental sounds). Auditory presentation may have exaggerated this. It is also important to note that the study by Peeters (2016) did find a smaller N400 in response to onomatopoeic vs. non-onomatopoeic words, which was interpreted as being indicative of facilitated word retrieval, and thus consistent with the present results.

We found some evidence that list context plays a role in the effects of iconicity. Indeed,

LDT reaction time and accuracy in the ELP were not affected by iconicity. Note that ELP participants would have received a much smaller proportion of iconic words than the participants we tested in Experiments 1 and 2. It may be that this greater proportion of iconic words served to

148 shift participants’ attention to iconicity. Indeed, the effect of iconicity on reaction time in the non- degraded LDT from Experiment 1 emerged over the course of that experiment. Thus, it seems that participants’ use of iconicity in LDT is somewhat strategic; they can shift their response strategy to rely more heavily on iconicity when it is beneficial to do so. This is consistent with other evidence for attentional control and strategic reliance on relevant lexical and semantic variables in LDT (Balota, Paul, & Spieler, 1999; Hargreaves & Pexman, 2012).

The results from Experiments 1 and 2 demonstrate that iconicity can confer an advantage in language processing, in addition to aiding in language learning (e.g., Imai et al., 2008) and increasing the vividness of communication (e.g., Lockwood & Tuomainen, 2015). Vocal iconicity has been also argued to have played a role in language evolution (Perlman et al., 2015;

Perlman & Lupyan, 2018). Various factors have been proposed to act on the relative balance of iconicity and arbitrariness in the lexicon over generations (for discussions of this topic see

Perniss & Vigliocco, 2014; Sidhu & Pexman, 2018a; Winter et al., 2017). One may speculate that the facilitatory role of iconicity in processing is one such factor, and that it might increase the chances of iconic forms being maintained over time.

In addition to testing for a benefit of iconicity, we also hoped to gain insight into how such an effect would fit within existing models of word recognition. We speculated that iconicity may have an effect via links from phonology to semantics that are special in some way (see also

Meteyard et al., 2015; Vigliocco & Kita, 2006; for evidence of phonemes activated cross modal information see Lockwood, Hagoort, & Dingemanse, 2016): iconic words could possess extra links (i.e., a quantitative difference) or links that are more direct in nature (i.e., a qualitative difference). To investigate this possibility, we examined whether increasing participants’ reliance on phonology by degrading visual stimuli (in Experiment 1) or having participants respond based

149 on phonology (Experiment 2) increased the effects of iconicity. We found no evidence that these manipulations increased effects of iconicity. It is worth noting however that, while non- significant (p = .06), the numerical trend that emerged when comparing the clear condition of

Experiment 1 with the phonological lexical decision task in Experiment 2, with iconicity playing a larger role in the latter.

The lack of a phonological effect speaks against the possibility that iconic words have extra links from phonology to semantics. Were this the case, increasing participants’ reliance on phonology should have allowed these extra links to facilitate processing. However, it still may be the case that iconic words have more direct links from phonology to semantics, and that these links facilitate processing regardless of the extent to which words are processed phonologically.

That is, while a null result with regards to the interaction between phonology and iconicity speaks against iconic words having quantitatively different links as compared to non-iconic words, it does not necessarily rule out the possibility that iconic words have qualitatively different links.

Additionally, recall that in Experiment 2, pseudohomophones based on words with a higher iconicity were responded to faster and more accurately. This suggests some role of phonology as correctly identifying a pseudohomophone as sounding like a word in the PLDT largely depends on the processing of phonology as the stimulus has no extant mapping from orthography to semantics.

Of course, it is also simply possible that effects of iconicity do not arise from special links between phonology and semantics. If so, how else might we account for them? One possibility is that the semantic representations of iconic words are special in some way. For instance, Meteyard et al. (2015) speculated that iconic words may have additional connections from semantic representations to modality-specific features. Indeed, previous work has shown relationships

150 between iconicity and sensory experience (Sidhu & Pexman, 2018b; Winter et al., 2017) and concreteness (Lupyan & Winter, 2018). As iconicity ratings become available for a greater number of items, future research should explore the role of various semantic dimensions in processing advantages of iconic words.

Conclusion

We found that iconicity provides a benefit to visual word recognition in typical adults.

Thus, iconicity contributes to language processing and should be considered in models of word recognition. On a larger scale, these findings demonstrate another benefit of iconicity to language.

151

Acknowledgements

The authors thank Mark Dingemanse and Darin Flynn for helpful correspondences about matters related to this work. The authors also thank Kristen Deschamps and Stella Heo for assistance running the experiments.

152

Chapter 5: Semantic Neighbourhood Density, Sensory Experience and Iconicity

Introduction

The nature of the relationship between the form of a word and its meaning has been debated since at least the time that Plato’s Cratylus takes place (5th century BC; Sedley, 2006).

The debate centers on whether the form of a word (i.e., its articulation and phonology) is related to its meaning. One position is that form and meaning have an arbitrary relationship, without any special connection (e.g., Hockett, 1963), such that aspects of a word’s form cannot be used as clues to its meaning (Dingemanse, Blasi, Lupyan, Christiansen, & Monaghan, 2015). For instance, consider the seemingly arbitrary word apple and the challenge of deriving its meaning based solely on its form.

However, it is also possible for aspects of a word’s form to be non-arbitrarily related to its meaning. One example of this is iconicity, in which aspects of a word’s form map onto aspects of its meaning (Emmorey, 2014; Taub, 2001). For instance, in the word ding aspects of form (i.e., abrupt onset, fading onset) map onto aspects of meaning via resemblance (Taub, 2001). While spoken languages (this article’s focus) allow for direct iconic mapping of auditory features, it is also possible for other kinds of sensory features to map onto form indirectly via sound symbolic associations. For instance, there is a well-documented association between the phonemes /b/, /m/,

/l/, /n/, /u/ and /o/, and roundness (i.e., the Maluma/Takete effect; Köhler, 1929). Thus a word like balloon, which denotes a round object, could be considered indirectly iconic. Blasi,

Wichmann, Hammarström, Stadler and Christiansen (2015) speculated that iconicity might explain their finding of consistency in the phonemes occurring in several basic vocabulary words, across nearly two-thirds of the world’s languages.

153

Instead of viewing the distinction between arbitrary and non-arbitrary words categorically, it has been suggested that words throughout the lexicon can have both arbitrary and non-arbitrary elements (e.g., Dingemanse et al., 2015). The word hiccups, for instance, sounds like the meaning it conveys (i.e., an iconic property), but its meaning cannot be derived purely from its form (i.e., an arbitrary property). By this view, the distinction between arbitrariness and non-arbitrariness is continuous, with non-arbitrariness (e.g., iconicity) appearing in varying amounts, throughout the lexicon.

Perry, Perlman, and Lupyan (2015) quantified this, by having participants rate the iconicity of 592 English words on a scale from -5 to 5: -5 indicated the word sounded opposite to its meaning, 0 indicated the word sounded nothing like its meaning, and 5 indicated the word sounded just like its meaning. Words were presented visually (Experiment 1) or auditorily

(Experiment 2). In both cases, words’ average rated iconicity was significantly higher than 0

(MExperiment 1 = 0.75, SDExperiment 1 = 0.99; MExperiment 2 = 0.78, SDExperiment 2 = 0.98), suggesting modest iconicity in this sample of English words (this was replicated in a larger item set by

Winter, Perlman, Perry, & Lupyan, 2017). The authors found, further, that onomatopoeic words were the most iconic, followed by interjections, adjectives and verbs. On average, nouns and function words were not rated as iconic. Lastly, Perry et al. (2015) also discovered that more iconic words tend to be acquired earlier.

These results suggest that iconicity may be present throughout the lexicon, but that it is not predominant. This is somewhat puzzling, given that iconicity makes communication more direct and vivid (Lockwood & Dingemanse, 2015), and facilitates language learning (for a review see Imai & Kita, 2014). If language is viewed from a cultural evolution standpoint, then features that improve its processing and learnability should survive and become more common

154

(Monaghan, Christiansen & Fitneva, 2011). Thus one might wonder why iconicity is not more prevalent in spoken languages.35

One possibility is that not all concepts can be easily mapped onto their word form. In examining Perry et al.’s (2015) ratings, it seems iconic mappings tend to require sensory features.

This is to be expected if we consider how iconic mappings take place. Direct mappings will mostly involve auditory sensations. In addition, indirect mappings will often involve sensory experience since most sound symbolic associations involve sensory features (for a review see

Lockwood & Dingemanse, 2015). Indeed, Winter et al. (2017) found that concepts rated as having a greater amount of associated sensory experience (SER; Juhasz & Yap, 2013) tended to have more iconic word forms. This was particularly true for concepts with a greater amount of auditory and tactile experience, suggesting that these features lend themselves well to iconic mappings in spoken languages (see also Dingemanse, 2012a).

However, there are likely to be other factors that determine the iconicity of a given concept’s word form.36 One relevant proposal is that while there are advantages to iconicity, there are also costs; namely, iconicity may lead to ambiguity (Gasser, 2004; Imai & Kita, 2014;

Monaghan et al., 2011; Monaghan et al., 2014). Consider the Maluma/Takete effect. If the naming of apples, oranges, and peaches were constrained by this form of iconicity, we might end up with these round fruits being named mulu, molo and lomo. Because similar meanings would beg similar forms, an iconic language would be populated by sets of words with similar forms

35 We do not wish to make the mistake of evaluating the iconicity of language as a whole based on spoken Indo European languages. Indeed many non-Indo European spoken languages contain many iconic words. However, even in these languages, we might ask why iconicity isn’t more prevalent. 36 Winter et al. (2017) also found that less frequent, less imageable words were more iconic. Imageability was not considered here, as it was only available for 63.32% of the words in our final sample.

155 and meanings. This would lead to ambiguity, and deficiencies in processing and learnability (e.g.,

Gasser, 2004).

However, concepts are not equally distributed in semantic space. On the contrary, some concepts have dense semantic neighbourhoods, in which there are many concepts with similar meanings, while others have sparse neighbourhoods. For instance, many concepts are similar in meaning to apple, while fewer are similar to balloon. This leads to the prediction that concepts with sparser semantic neighbourhoods can afford to have more iconic word forms, and enjoy the benefits of iconicity, without risking confusion. Conversely, concepts with denser semantic neighbourhoods may need to have relatively more arbitrary word forms. This illustrates how iconicity and arbitrariness can play complimentary roles in language (see Dingemanse et al.,

2015; Perniss & Vigliocco, 2014), with each taking on a more prominent role in different contexts.

The main goal of the present paper was to examine the possibility that words used to describe concepts with sparser semantic neighbourhoods will be relatively more iconic. In addition, we sought to further examine the relationship between sensory experience and iconicity described in Winter et al. (2017). Because earlier acquired concepts tend to be both richer in sensory experience (Juhasz, Yap, Dicke, & Gullick, 2011) and more iconic (Perry et al., 2015), it is important to examine the alternate explanation that the relationship between SER and iconicity is attributable to age of acquisition. We also examined whether the effects of semantic neighborhood density and SER are additive. To accomplish these goals, we used the ratings collected by Perry et al. (2015) and Winter et al. (2017) to quantify iconicity. We used Shaoul and Westbury’s (2010) average radius of co-occurrence (ARC) variable to measure semantic neighbourhood density. ARC uses lexical co-occurrence information to quantify semantic

156 similarity between a word and its neighbours in semantic space. We also used Juhasz and Yap’s

(2013) sensory experience ratings (SER), to quantify the amount of sensory information associated with a concept.

Method

Materials and Procedure

In these analyses we used the combined iconicity ratings from Perry et al. (Experiment 1;

2015) and Winter et al. (2017). Our main interest was quantifying the iconicity of each word on a spectrum from arbitrary to non-arbitrary (i.e., iconic), so we eliminated 72 words that an item analysis revealed had iconicity ratings significantly below zero (α = .1). Words rated below zero in the iconicity ratings studies were judged to have forms that mapped onto the opposite of their meaning, and thus were not arbitrary. Our interest was in the positive ratings since these captured degrees of iconicity, with lower values indicating more arbitrariness. We used a liberal criterion here because the danger of a Type 1 error (i.e., excluding words whose ratings were not actually different than zero) was less problematic than including words that affected the validity of the scale. In addition, we removed six onomatopoeic words and six interjections, since our goal was to examine the factors determining iconicity in the general lexicon.

Our analysis consisted of a hierarchical multiple regression predicting these iconicity ratings. In Step 1, we included control variables that are often used to control for lexical factors

(e.g., Yap et al., 2012): letter length, number of phonemes, word frequency (Brysbaert & New,

2009), and orthographic Levenshtein distance (Yarkoni, Balota, & Yap, 2008). We also included age of acquisition (AoA; Kuperman, Stadthagen-Gonzalez, & Brysbaert, 2012). Lastly, as iconicity has been shown to vary by word type, we included three dummy coded predictors for words’ status as adjective/adverbs, verbs, or nouns. In Step 2 we added SER (Juhasz & Yap,

157

2013), and in Step 3 we added ARC (Shaoul & Westbury, 2010). Larger SER values indicate greater associated sensory experience; larger ARC values indicate denser semantic neighbourhoods. Finally, in Step 4 we added an interaction between SER and ARC. Data on all dimensions were available for 1709 words.

Table 22

Correlations among variables.

Variable 1 2 3 4 5 6 7

1. Iconicity ---

2. Length -.05* ---

3. Number of Phonemes -.06* .81*** ---

4. Frequency -.21*** -.31*** -.32*** ---

5. OLD -.08** .81*** .71*** -.28*** ---

6. AoA .02 .29*** .32*** -.59*** .28*** ---

7. SER .18*** .30*** .25*** -.24*** .27*** -.16*** ---

8. ARC -.34*** -.16*** -.14*** .73*** -.17*** -.37*** -.21***

Note. OLD = Orthographic Levenshtein distance; AoA = Age of Acquisition; SER = Sensory

Experience Rating; ARC = Semantic Neighbourhood Density

* p < .05; ** p < .01; *** p < .001

Results

The results from Step 1 showed that less frequent, less orthographically distinct and earlier acquired words were rated as more iconic. Replicating previous findings, nouns were also

158 found to be less iconic than other word types. More importantly, the results from Steps 2 and 3 showed that words associated with more sensory experience, and with sparser semantic neighbourhoods, tended to be more iconic. The ΔR2 values reveal that both of these semantic

2 2 variables accounted for incremental variance: sr SER = .03 and sr ARC = .04; these values are the same if ARC is entered before SER. See Table 23 for a summary of the model. Zero-order correlations indicated that SER and ARC were each associated with iconicity, when analyzed separately for adjectives/adverbs, verbs and nouns (see Table 24 and Figure 9). Lastly, there was also a significant interaction between SER and ARC (see Figure 10). Investigating the regions of significance indicated that higher SER was positively related with iconicity, except for words with especially high ARC values (z > 1.63).

159

Table 23

Results of hierarchical regression predicting iconicity.

Variable B SEB β sr2 R2 ΔR2

Step 1 (Control variables) .14*** .14***

Length 0.03 0.04 0.04 0.00

Number of Phonemes -0.06 0.04 -0.06 0.00

Frequency -0.48 0.04 -0.42 0.09***

OLD -0.22 0.09 -0.10 0.003*

AoA -0.08 0.01 -0.18 0.02***

Adjective/Adverb -0.00 0.16 -0.00 0.00

Verb -0.02 0.15 -0.01 0.00

Noun -0.57 0.15 -0.27 0.01***

Step 2 .17*** .03***

SER 0.22 0.03 0.22 0.03***

Step 3 .21*** .04***

ARC -2.53 0.28 -0.30 0.04***

Step 4

SER x ARC Interaction -0.54 0.20 -0.06 0.003** .21*** .003**

Note. OLD = Orthographic Levenshtein distance; AoA = Age of Acquisition; SER = Sensory

Experience Rating; ARC = Semantic Neighbourhood Density

* p < .05; ** p < .01; *** p < .001

160

Table 24

Correlations between iconicity and ARC and SER, for each word type.

Correlation with Iconicity

Word Type n SER ARC

Adjectives/Adverbs 219 .18** -22**

Function Words 57 .12 -.12

Verbs 362 .46*** -.48***

Nouns 1071 .13*** -.30***

Note. SER = Sensory Experience Rating; ARC = Semantic Neighbourhood Density

* p < .05; ** p < .01; *** p < .001

161

Figure 9. The moderating effect of ARC on SER’s relationship with iconicity. While concepts with higher SER values tend to have more iconic word forms, this is tempered in dense semantic neighbourhoods.

162

Figure 10. Scatterplots show the relationships between SER/ARC and iconicity, separately for adjectives/adverbs, verbs and nouns. Gray lines represent LOWESS functions (created using lowess function in R with the default smoother value of 66.67%), black lines represent zero-order correlations.

Replicating previous research, we found that nouns were significantly less likely to be iconic than other word types. This was confirmed by an independent samples t-test comparing the iconicity of nouns (M = 0.73, SD = 0.97) to that of adjectives/adverbs/verbs (M = 1.17, SD =

1.10), t(1076.22) = 8.08, Glass’s ΔAdjectives/Adverbs/Verbs = 0.40, p < .001 (unequal variances). We conducted supplementary analyses to further explore this finding. Surprisingly, nouns had

163 equivalent semantic neighbourhood densities (M = 0.56, SD = 0.12) as adjectives/adverbs/verbs

(M = 0.56, SD = 0.14), t(1017.70) = 0.02, p = .99 (unequal variances). Thus, this factor cannot explain difference in iconicity between word types. Additionally, nouns were associated with more sensory experience (M = 3.41, SD = 0.98) than were adjectives/adverbs/verbs (M = 3.03,

SD = 0.94), t(1650) = 7.52, Cohen’s d = 0.39, p < .001. However, recall that Winter et al. (2017) discovered that certain kinds of sensory experience (i.e., auditory and tactile) were more important to iconicity than others. As such, it is interesting to note that nouns were associated with less auditory sensory experience (Lynott & Connell, 2009; 2013; Winter, 2016; M = 0.53,

SD = 0.95) than were adjectives/adverbs/verbs (M = 1.08, SD = 1.33), t(420.13) = 5.27, Glass’s

ΔAdjectives/Adverbs/Verbs = 0.42, p < .001 (unequal variances). Nouns were not, however, associated with less tactile sensory experience (M = 0.82, SD = 0.92) than adjectives/adverbs/verbs (M =

0.85, SD = 1.12), t(454.25) = 0.28, p = .78 (unequal variances). Nevertheless, this is evidence that while nouns may be associated with more sensory experience overall, adjectives/adverbs/verbs may be associated with the particular kind of experience that lends itself to an iconic mapping.

Discussion

In recent research, arbitrariness and iconicity have been shown to exist on a spectrum, with words varying in their arbitrary and iconic elements (e.g., Perry et al., 2015). We examined the prediction that concepts with sparser semantic neighbourhoods could afford to have more iconic forms, and further investigated the finding that concepts evoking a greater amount of sensory information are more mappable and thus more iconic (Winter et al., 2017).

We found evidence for both predictions. Even after accounting for a variety of important lexical-semantic variables, words’ sensory experience ratings and semantic neighbourhood density were related to their iconicity. This former finding replicates Winter et al. (2017) even

164 after accounting here for effects of age of acquisition, demonstrating that the relationship between sensory experience and iconicity cannot be fully attributed to this third variable. We also found further support for the notion that certain types of sensory experience are especially important to iconic mappings. While the relatively less iconic syntactic class of nouns was associated with more sensory experience than adjectives/adverbs/verbs, this latter syntactic group was associated with a greater amount of auditory sensory experience, a type of sensory feature that lends itself to iconic mappings in spoken language (Dingemanse, 2012a; Winter et al., 2017).

The unique contribution of the present paper is demonstrating that semantic neighbourhood density is an important factor in the iconicity of a given concept’s word form.

This factor explained as much, or more, unique variance as: SER, frequency, and AoA, factors that have all previously been shown to be important to iconicity. The effect of semantic neighbourhood density highlights the cooperative roles of iconicity and arbitrariness in language

(see Dingemanse et al., 2015; Perniss & Vigliocco, 2014). While iconicity conveys benefits to language, and can play a large role in some concepts’ word forms, there are instances in which arbitrariness must play the larger role, to avoid ambiguity and aid in discriminability (e.g., for concepts with dense semantic neighbourhoods).

The relationship between semantic neighbourhood density and iconicity has relevance for our understanding of the acquisition and evolution of iconic words. Some have suggested that early-acquired language can afford to be more iconic because early language maps sparser semantic space (Gasser, 2004). Others have theorized that the original forms of many words may have been iconic, but that forms became more arbitrary as more words were added and semantic neighbourhoods became denser (Imai & Kita, 2014). Our results are consistent with both of these claims. We also found a significant interaction between SER and ARC, the nature of which is

165 consistent with these theories. In general, concepts are more able to have iconic word forms if they have a greater amount of sensory information. However, once a given region of semantic space becomes cluttered, words for these concepts tend to become arbitrary, regardless of their associated sensory experience. Instead of being additive in every case, high semantic neighbourhood density seems to nullify the effects of sensory experience.

The present study provides new insight about factors that modulate the iconicity of word forms and thus advances the ancient debate about the relationship between a word’s form and its meaning. We found that semantic neighbourhood density is important to iconicity, and that it moderates the extent to which sensory experience predicts the iconicity of concepts’ word forms.

Of course, each of these variables explained a small amount of overall variance in iconicity, suggesting that there are likely other factors that are as yet unidentified. Nevertheless, these results reinforce the notion that iconicity is not simply a linguistic oddity but rather a more general property of the lexicon, one that cooperates with arbitrariness in shaping language.

166

Acknowledgements

The authors would like to thank Bodo Winter and Lynn Perry for providing trial level data of their iconicity ratings. This research was supported by the Natural Sciences and

Engineering Research Council of Canada (NSERC) through a Postgraduate Scholarship to DMS

[CGSD3 476111 2015] and a Discovery Grant to PMP [RGPIN 217309-2013]; and by Alberta

Innovates: Health Solutions (AIHS) through a Graduate Scholarship to DMS [201500125-1 CA#

3874].

167

Chapter 6: Conclusion

In Chapter 2 I synthesized the existing literature and suggested there are five potential mechanisms by which phonemes may become associated with perceptual/semantic features. It is my hope that this facilitates future experiments on the existence of these mechanisms. As mentioned in the chapter, I don’t expect any one of these mechanisms to explain all of sound symbolism. Different mechanisms may be involved in different associations. Additionally, multiple mechanisms may combine in the creation of a single association.

In Chapter 3 I presented a novel sound symbolic association between phonemes in real first names and personality factors from the HEXACO. One of the main conclusions that can be drawn from these studies is that sound symbolic associations can also exist for abstract dimensions like personality. This broadens the scope of sound symbolism and, more notably, the types of effects that the mechanism of sound symbolism must be able to explain. I ruled out some possible mechanisms for this effect and speculated as to how some of the mechanisms described in Chapter 2 might have been able to explain these results. Additionally, these studies demonstrated that sound symbolic associations can have an effect in existing language.

In Chapter 4 I turned my attention to iconic relationships in existing language. I found that the processing of iconic language is facilitated (vs. arbitrary language), with responses to these words being faster and more accurate. While we theorized that this may be due to special links between phonology and semantics, the data were inconclusive with regards to this question.

Nevertheless, these results demonstrated a novel benefit of iconicity.

In Chapter 5 I explored the question of what kinds of meanings tend to have iconic words in English. The results suggested that more unique meanings, and those with a greater amount of associated sensory information, were most likely to have iconic forms. This supports the proposal

168 that iconicity can be facilitatory in sparser semantic space, but that it could potentially be problematic in crowded semantic space in which similar meanings with similar forms could lead to confusion. In addition, it seems that sensory information may be key to establishing an iconic mapping. These two results speak to the question of why iconicity isn’t more prevalent in

English.

Limitations

In each chapter I discussed limitations specific to the experiments contained therein.

However, there are also some overarching limitations. The first is that all of the experiments were quite artificial in nature. That is, they involved participants responding to stimuli presented in isolation, and responding in a constrained way (i.e., with a rating scale or dichotomous decision).

This is quite an impoverished way of examining sound symbolism and iconicity, especially considering that one of the functions of iconicity is to make communication more vivid

(Lockwood & Dingemanse, 2015). It may be that there are aspects of sound symbolic and iconic processing that are not being tapped when using such an approach. Alternatively, the contexts of the tasks I used may have directed participants’ attention towards sound symbolism and iconicity.

For example, the studies in Chapter 3 asked participants to judge how names went along with a specific target, perhaps encouraging participants to consider potential sound symbolic links with that target. In Chapter 4, the lexical decision task contained a greater proportion of iconic words than does language in general. This may have cued participants to pay attention to these iconic links. Thus, it is possible that these tasks overestimate the effects of sound symbolism and iconicity.

Another limitation is that all experiments focused on English stimuli and speakers. As reviewed in Chapter 1, there is reason to believe that sound symbolism effects can be moderated

169 by the phonetic inventory of one’s language (Styles & Gawne, 2017). Additionally, it appears that some languages contain a greater proportion of iconic words than others. For instance,

English is largely devoid of ideophones, which are prominent in many other languages (see

Perniss et al., 2010). This could conceivably affect how iconic language is processed.

Nevertheless, the available evidence seems to suggest that native language serves as a moderator of sound symbolism effects (e.g., Styles & Gawne, 2017). Thus, conclusions drawn from a sample of English speakers need not be invalidated or treated with suspicion. Rather, the concern is that they may not provide a sense of the effect in its entirety.

It is important to point out that in Chapters 4 and 5 I operationalized iconicity using subjective measures. Further, this was based on only an estimated 17 ratings per word.37 Of course subjective ratings are common in the field (e.g., subjective ratings of emotion; Warriner,

Kuperman, & Brysbaert, 2013) and are not in themselves problematic. However, one might raise the concern that judging the fit between a form and a meaning is a more difficult task than judging the valence of a word, and potentially more open to interpretation. This could introduce discrepancies in the ways that different participants approached the task. Attempts have been made to compute a measure of acoustic similarity among the phonemes in onomatopoeia and the sounds to which they refer (Assaneo, Nichols, & Trevisan, 2011). This might present an objective way of quantifying direct iconicity. Unfortunately, indirect iconicity may only be quantifiable using subjective measures, since it is by its very nature a subjective experience (see Hutton,

1989).

Finally, iconicity ratings are only available for a small number of words (3,001 at present). This pales in comparison to, for instance, datasets of emotion ratings (13,915; Warriner

37 This is estimated based on the stated number of participants, words and words rated per participant, in Winter et al. (2017).

170 et al., 2013) or concreteness ratings (37,058; Brysbaert, Warriner, & Kuperman, 2014). This limits the scope of research that can be conducted, as the small set of words will have restricted overlap with other existing rating sets and will create difficulties when matching items for experiments. The small set of items also creates the possibility that any effects observed are limited to this specific sample of English words rather than generalizing to the lexicon more broadly.

Future Research

Sound Symbolism

As reviewed at the close of Chapter 2, there is a need for future research to directly test the five suggested mechanisms of sound symbolic association. This could occur in several ways.

One could hypothesize the existence of associations given certain mechanisms, and then test for them. Another approach could be to attempt to create novel associations using a particular mechanism. It is likely that no single mechanism explains all of sound symbolism. Thus, instead of explaining sound symbolism in general, a more prudent approach might be to choose a single association and attempt to unpack the interplay of multiple mechanisms in its creation. This could involve testing participants at different ages to observe how associations change over time, and correspond with developmental milestones.

One particular topic that deserves further study is the fact that many sound symbolic associations seem to cluster together. For instance, there are a constellation of dimensions associated with front/back vowels (e.g., brightness, smallness, quickness). There were also multiple personality factors that were all associated with sonorant names in Chapter 3.

Understanding what causes this could be very informative in the search for an underlying mechanism. Future work might take a big data approach and explore the associations between

171 phonemes and multiple dimensions at once (e.g., Monaghan & Fletcher, 2019; Tzeng et al., 2016;

Westbury et al., 2018), with the goal of exploring the interplay of these different associations.

This could reveal that a single dimension mediates various associations, or that the associations are determined at a higher order of abstractness (i.e., based on some higher order feature that the dimensions share).

Iconicity

In Chapters 4 and 5 I treated iconicity as a homogenous concept. In actuality there are different types of iconicity in the lexicon. A notable distinction is that between direct and indirect iconicity (or onomatopoeic and non-onomatopoeic iconicity). Since these are so different in nature, it stands to reason that they could lead to different effects. Thus, future work should attempt to distinguish between these kinds of iconicity and explore the sorts of effects discussed in Chapters 4 and 5 separately for each. Indeed, indirect links between phonology and semantics may require mediation, potentially leading to a different pattern of results on word processing tasks such as an LDT. In addition, indirectly iconic words may be associated with different kinds of sensory experience than directly iconic words (since they are not restricted to auditory phenomena). A dimension that may be key in distinguishing between these two kinds of iconicity is imitativeness: the extent to which a word imitates its meaning. We would expect directly iconic words to score high on this dimension but indirectly iconic words to score lower.

The results presented in Chapter 5 suggested that words in crowded areas of semantic space are prevented from being iconic due to the possibility of confusion. I discussed this as a diachronic effect (i.e., taking place over time), but only by using synchronic data (i.e., a snapshot of language at a given moment). A major topic requiring future attention is how iconicity emerges and changes over time in the lexicon. In terms of how iconic forms emerge in language,

172 it’s possible that iconic words are coined specifically with an iconic mapping in mind (see Imai &

Kita, 2014), and then persevere in language, protected from language change (see Nuckolls,

1999). It would be interesting to examine how the effects observed in Chapter 5 play out over time. For instance, one would make the prediction that as iconic words with a similar meaning to iconic word W enter the language, W might become less iconic.

Finally, as mentioned, the work here has used very constrained laboratory tasks. This may not capture the full scope of iconicity’s effects on language. Indeed, one proposed benefit of iconicity is that it makes communication more vivid (Lockwood & Dingemanse, 2015). As such, future tasks should explore the effects of iconicity (and sound symbolism) on communication.

Rather than using isolated word processing tasks, researchers might, for instance, use tasks involving full sentences, and pairs of participants communicating with one another (e.g., Verhoef,

Roberts, & Dingemanse, 2015).

Theoretical Contributions

The specific, novel contributions of this work are as follows:

a) Sound symbolic associations exist between phonemes and the abstract dimension of

personality. This is true even in the context of existing names.

b) The processing of iconic words is facilitated relative to arbitrary words, at least in certain

task contexts.

c) Iconic words tend to have more unique meanings, and more associated sensory

experience, than arbitrary words.

In addition to these, I see this work as making several broad advancements to the literature. One is a shift towards examining the mechanisms underlying sound symbolism. The maluma/takete effect was first written about in 1929, and since then there have been numerous

173 demonstrations of it and similar effects. These have been extremely informative as they have enriched our understanding of the nuances of the effect and the conditions under which it will emerge. However, the field is now in a good place to begin asking why these effects exist at all. It is my sincere hope that Chapter 2 facilitates this investigation, with studies aimed at disentangling the five possible mechanisms. Additionally, when new sound symbolic associations are demonstrated (as in Chapter 3), their discussion must include an exploration of their underlying mechanism, and perhaps a discussion of insights they might offer regarding the mechanisms of sound symbolism in general.

Further, many of the sound symbolic effects demonstrated since 1929 have involved perceptual dimensions. It is notable that Chapter 3 demonstrated a more abstract form of sound symbolism. I believe this to be important as it might rekindle interest in the connotative dimensions of sound symbolism, a topic that was well studied in the 1960’s and 1970’s (e.g.,

Greenberg & Jenkins, 1966; Miron, 1961). As covered in Chapter 2, I believe that these associations, and their interrelationships, might be key to understanding the mechanism of sound symbolism.

Another step forward is examining sound symbolism and iconicity in broad, existing language. Most studies of sound symbolism have involved nonwords. As covered in Chapter 3, there are reasons to expect sound symbolism effects will not emerge in existing language.

Chapter 3 demonstrated that effects will still emerge with existing language, though they may be attenuated. In addition, many of the studies of iconicity have exclusively involved onomatopoeia and/or ideophones. I went beyond that in Chapter 4 by including non-onomatopoeic iconic words. I excluded onomatopoeic words in Chapter 5 and instead surveyed a broad selection of the

174 lexicon. All of the effects demonstrated in these chapters suggest that sound symbolism and iconicity are much more broadly relevant than previous studies may have suggested.

Finally, it has been my impression that work on sound symbolism and iconicity has existed in isolation from work on the psychology of language more broadly (though see Meteyard et al., 2015; Reilly et al., 2012; Westbury, 2005; Winter et al., 2018). Now that sound symbolism and iconicity are established phenomena, their relationship with the psychology of language should be investigated. In Chapter 4 I considered effects of iconicity on a common task in word processing research (i.e., the LDT), and how iconicity might fit in to existing models of word processing. In Chapter 5 I examined how iconicity related to several important dimensions of language. I hope that these investigations help to bring sound symbolism and iconicity into the mainstream.

Future Theoretical Refinements

Sound Symbolism

In Chapter 1 I operationalized sound symbolism as phonemes activating associated perceptual and/or semantic features. This is only a best guess. It is true that after all this time, the exact cognitive process underlying sound symbolism is not entirely understood. Exactly what it means for /m/ to be associated with roundness requires refinement. Is it due to some component feature of the phoneme, or the phoneme as a whole? Is “roundness” the sort of feature that is implicated in the meaning of ball, or is it rather a non-lexical feature? Instead of a phoneme activating a feature, might it simply be compatible with a feature? The field should work to clarify all of these points, to arrive at a more precise definition of sound symbolism. These clarifications will likely also contribute to a better understanding of the mechanisms creating these associations (and vice versa).

175

Iconicity

As alluded to earlier, the concept of iconicity is multifaceted. It would be beneficial for the field to use specific terminology when referring to the phenomenon to distinguish between, for instance, direct and indirect forms of iconicity. This would ideally lead to theoretical advancements that investigate either phenomena separately. They may indeed prove to be distinct phenomena. Regarding indirect iconicity, there is still uncertainty regarding its nature (e.g., is a comparison among contrasting forms and meanings required; see Dingemanse et al., 2015).

Future work should clarify the extent to which these are separate phenomena.

There are several concepts related to iconicity that should also be considered in future theorizing. One is the notion of coercion (see Dingemanse, 2012b): the idea that certain instances of iconicity are only obvious in a given context. That is, certain form-meaning mappings may require a person to be looking for the mapping for it to be apparent. This is relevant to the concern that certain iconicity effects may be embellished by task context. Another is the dimension of imitativeness: the extent to which a form imitates its meaning. The interplay between these ideas and iconicity should be considered in future theories.

Conclusion

Humans tend to make associations across modalities (see Spence, 2011). Because language makes use of perceptual stimuli, this propensity to associate crossmodally brings a host of associations to the experience of language. Crossmodal properties piggyback on sounds and sensations that we use to communicate; they add another dimension of meaning to language, can facilitate its processing, and affect its structure.

Italo Calvino said: “The struggle of literature is in fact a struggle to escape from the confines of language; it stretches out from the utmost limits of what can be said; what stirs

176 literature is the call and attraction of what is not in the dictionary.” (1986, p. 16). In light of sound symbolism, we might say that words already have one (small) foot outside the confines of language. That they have always evoked properties beyond the ones to which they refer in the dictionary.

177

References

Akita, K. (2011). An embodied semantic analysis of psychological mimetics in Japanese.

Linguistics, 48, 1195–1220.

Allport, G. W., & Odbert, H. S. (1936). Trait-names: A psycho-lexical study. Psychological

Monographs, 47, 1–38.

Anderson, S. R. (1985). Phonology in the twentieth century: Theories of rules and theories of

representations. Chicago, IL: The University of Chicago Press.

Anderson-Clark, T. N., Green, R. J., & Henley, T. B. (2008). The relationship between first

names and teacher expectations for achievement motivation. Journal of Language and

Social Psychology, 27, 94–99.

Arbib, M.A. (2005). From monkey-like action recognition to human language: An evolutionary

framework for neurolinguistics. Behavioral and Brain Sciences, 28, 105–124.

Asano, M., Imai, M., Kita, S., Kitajo, K., Okada, H., & Thierry, G. (2015). Sound symbolism

scaffolds language development in preverbal infants. Cortex, 63, 196–205.

Ashton, M. C., & Lee, K. (2007). Empirical, theoretical, and practical advantages of the

HEXACO model of personality structure. Personality and Social Psychology Review, 11,

150–166.

Ashton, M. C., Lee, K., & Goldberg, L. R. (2004). A hierarchical analysis of 1,710 English

personality-descriptive adjectives. Journal of Personality and Social Psychology, 87,

707–721.

Assaneo, M. F., Nichols, J. I., & Trevisan, M. A. (2011). The anatomy of onomatopoeia. PloS

ONE, 6. doi:10.1371/journal.pone.0028317

178

Auracher, J. (2017). Sound iconicity of abstract concepts: Place of articulation is implicitly

associated with abstract concepts of size and social dominance. PLOS ONE, 12. doi:

10.1371/journal.pone.0187196

Aveyard, M. E. (2012). Some consonants sound curvy: Effects of sound symbolism on object

recognition. Memory & Cognition, 40, 83–92.

Baayen, H., Bates, D., Kliegle, R., & Vasishth, S. (2015). RePsychLing: Data sets from

Psychology and Linguistics experiments. R package version 0.0.4.

Baier, B., Kleinschmidt, A., & Müller, N. G. (2006). Cross-modal processing in early visual and

auditory cortices depends on expected statistical relationship of multisensory

information. Journal of Neuroscience, 26, 12260–12265.

Balota, D. A., Paul, S., & Spieler, D. H. (1999). Attentional control of lexical processing

pathways during word recognition and reading. In S. Garrod & M. Pickering (Eds.),

Language Processing, (pp. 15–57). East Sussex, UK: Psychology Press Ltd.

Balota, D. A., Yap, M. J., Hutchison, K. A., Cortese, M. J., Kessler, B., Loftis, B., …& Treiman,

R. (2007). The English lexicon project. Behavior Research Methods, 39, 445–459.

Bankieris, K., & Simner, J. (2015). What is the link between synaesthesia and sound symbolism?

Cognition, 136, 186–195. doi:10.1016/j.cognition.2014.11.013

Bar, M., & Neta, M. (2006). Humans prefer curved visual objects. Psychological Science, 17,

645–648.

Barr, D. J., Levy, R., Scheepers, C., & Tily, H. J. (2013). Random effects structure for

confirmatory hypothesis testing: Keep it maximal. Journal of Memory and Language, 68,

255–278.

Bates, D., Kliegl, R., Vasishth, S., & Baayen, H. (2015). Parsimonious mixed models.

179

arXiv:1506.04967 [stat]. Retrieved from http://arxiv.org/abs/1506.04967

Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models

using lme4. Journal of Statistical Software, 67, 1–48.

Bergen, B. K. (2004). The psychological reality of phonaesthemes. Language, 80, 290–311.

Berlin, B. (1994). Evidence for pervasive synesthetic sound symbolism in ethnozoological

nomenclature. In L. Hinton, J. Nichols, & J. Ohala (Eds.), Sound symbolism (pp. 76–93).

New York, NY: Cambridge University Press.

Bertrand, M., & Mullainathan, S. (2004). Are Emily and Greg more employable than Lakisha and

Jamal? A field experiment on labor market discrimination. American Economic Review,

94, 991–1013.

Best, C. T. 1995. A direct realist perspective on cross-language speech perception. In W. Strange

(Ed.), Speech perception and linguistic experience: Theoretical and methodological

issues in cross-language speech research (pp. 167–200). Timonium, MD: York Press.

Blasi, D. E., Wichmann, S., Hammarström, H., Stadler, P. F., & Christiansen, M. H. (2016).

Sound–meaning association biases evidenced across thousands of languages. Proceedings

of the National Academy of Sciences, 113, 1001.

Bolinger, D. (1950). Rime, assonance, and morpheme analysis. Word, 6, 117–36.

Borges, J. L. (1980). Seven Nights. New York, NY: New Directions Publishing Corporation.

Bottini, R., Barilari, M., & Collignon, O. (2019). Sound symbolism in sighted and blind. The role

of vision and orthography in sound-shape correspondences. Cognition, 185, 62–70.

Bowling, D. L., Garcia, M., Dunn, J. C., Ruprecht, R., Stewart, A., Frommolt, K. H., & Fitch, W.

T. (2017). Body size and vocalization in primates and carnivores. Scientific Reports, 7.

doi: 10.1038/srep41070

180

Bozzi, P., & Flores D’Arcais, G. B. (1967). Experimental research on the intermodal

relationships between expressive qualities. Archivio di Psicologia, Neurologia e

Psichiatria, 28, 377–420.

Bremner, A. J., Caparos, S., Davidoff, J., de Fockert, J., Linnell, K. J., & Spence, C. (2013).

“Bouba” and “ Kiki” in Namibia? A remote culture make similar shape-sound matches,

but different shape-taste matches to Westerners. Cognition, 126, 165–172.

doi:10.1016/j.cognition.2012.09.007

Brown, R. W., Black, A. H., & Horowitz, A. E. (1955). Phonetic symbolism in natural

languages. The Journal of Abnormal and Social Psychology, 50, 388–393.

Brysbaert, M., & New, B. (2009). Moving beyond Kučera and Francis: A critical evaluation of

current word frequency norms and the introduction of a new and improved word

frequency measure for American English. Behavior Research Methods, 41, 977–990.

Brysbaert, M., Warriner, A. B., & Kuperman, V. (2014). Concreteness ratings for 40 thousand

generally known English word lemmas. Behavior Research Methods, 46, 904–911.

Bürkner, P. (2017). brms: An R package for bayesian multilevel models using Stan. Journal of

Statistical Software, 80, 1–28.

Calvino, I. (1978). Invisible cities. New York: Harcourt Brace & Company.

Calvino, I. (1986). The uses of literature. New York: Harcourt Brace & Company.

Cassidy, K. W., & Kelly, M. H. (1991). Phonological information for grammatical category

assignments. Journal of Memory and Language, 30, 348–369.

Charlton, B. D., & Reby, D. (2016). The evolution of acoustic size exaggeration in terrestrial

mammals. Nature Communications, 7, 12739.

Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: a dual route

181

cascaded model of visual word recognition and reading aloud. Psychological Review, 108,

204–256.

Cottrell, C. A., & Neuberg, S. L. (2005). Different emotional reactions to different groups: a

sociofunctional threat-based approach to "prejudice". Journal of Personality and Social

Psychology, 88, 770–789.

Cowles, J. T. (1935). An experimental study of the pairing of certain auditory and visual

stimuli. Journal of Experimental Psychology, 18, 461–469.

Cuskley, C. (2013). Mappings between linguistic sound and motion. Public Journal of

Semiotics, 5, 39–62.

Cuskley, C. and Kirby, S. (2013). Synaesthesia, cross-modality, and language evolution. In J.

Simner, & E. Hubbard (Eds.), Oxford Handbook of Synaesthesia (pp. 869–907). Oxford,

UK: Oxford University Press.

Cuskley, C., Simner, J., & Kirby, S. (2015). Phonological and orthographic influences in the

bouba-kiki effect. Psychological Research. doi:10.1007/s00426-015-0709-2

Davies, N. B., & Halliday, T. R. (1978). Deep croaks and fighting assessment in toads Bufo

bufo. Nature, 274, 683–685.

Davis, R. (1961). The fitness of names to drawings. A cross-cultural study in Tanganyika. British

Journal of Psychology, 52, 259–268.

Deroy, O., Crisinel, A. S., & Spence, C. (2013). Crossmodal correspondences between odors and

contingent features: odors, musical notes, and geometrical shapes. Psychonomic Bulletin

& Review, 20, 878–896.

Deroy, O., and Auvray, M. (2013). A new Molyneux's problem: sounds, shapes and arbitrary

crossmodal correspondences. In O. Kutz, M. Bhatt, S. Borgo, & P. Santos (Eds.),

182

Proceedings of the Second Interdisciplinary Workshop The Shape of Things (SHAPES

2013), Rio de Janeiro, Brazil, April 3-4, 2013 . CEUR Workshop Proceedings, 1007, 61–

70. Retrieved from http://ceur-ws.org/Vol-1007/paper3.pdf

Dingemanse, M. (2012a). Advances in the cross-linguistic study of ideophones. Linguistics and

Language Compass, 6, 654–672. doi:10.1002/lnc3.361

Dingemanse, M. (2012b). Coerced iconicity in writing and speech. Retrieved from

https://semioticon.com/semiotix/2012/07/coerced-iconicity-in-writing-and-speech/

Dingemanse, M., Blasi, D. E., Lupyan, G., Christiansen, M. H., & Monaghan, P. (2015).

Arbitrariness, iconicity, and systematicity in language. Trends in Cognitive Sciences, 19,

603–615. doi:10.1016/j.tics.2015.07.013

Dingemanse, M., Schuerman, W., Reinisch, E., Tufvesson, S., & Mitterer, H. (2016). What sound

symbolism can and cannot do: testing the iconicity of ideophones from five

languages. Language, 92, e67–e83.

Dolscheid, S., Shayan, S., Majid, A., & Casasanto, D. (2013). The thickness of musical pitch:

Psychophysical evidence for linguistic relativity. Psychological Science, 24, 613–621.

D’Onofrio, A. (2013). Phonetic detail and dimensionality in sound-shape correspondences:

Refining the bouba-kiki paradigm. Language and Speech, 57, 367–393.

doi:10.1177/0023830913507694

Drellishak, S. (2007, January). Statistical techniques for detecting and validating phonesthemes.

Paper presented at the Linguistics Society of America Annual Meeting, Anaheim, CA.

Drijvers, L., Zaadnoordijk, L., & Dingemanse, M. (2015). Sound-symbolism is disrupted in

dyslexia: Implications for the role of cross-modal abstraction processes. In Proceedings of

the 37th Annual Meeting of the Cognitive Science Society (pp. 602–607). Austin, TX:

183

Cognitive Science Society.

Eberhardt, M. (1940). A study of phonetic symbolism of deaf children. Psychological

Monographs, 52, 23–41.

Elsen, H. (2017). The two meanings of sound symbolism. Open Linguistics, 3, 491–499.

Emmorey, K. (2014). Iconicity as structure mapping. Philosophical Transactions of the Royal

Society of London. Series B, Biological Sciences, 369, 20130301.

doi:10.1098/rstb.2013.0301

Ernst, O. M. (2007). Learning to integrate arbitrary signals from vision and touch. Journal of

Vision, 7, 1–14. doi:10.1167/7.5.7

Evans, K. K., & Treisman, A. (2010). Natural cross-modal mappings between visual and auditory

features. Journal of Vision, 10, 1–12.

Evans, S., Neave, N., & Wakelin, D. (2006). Relationships between vocal characteristics and

body size and shape in human males: an evolutionary explanation for a deep male

voice. Biological Psychology, 72, 160–163.

Fadiga, L., Craighero, L., Buccino, G., & Rizzolatti, G. (2002). Speech listening specifically

modulates the excitability of tongue muscles: a TMS study. European Journal of

Neuroscience, 15, 399–402.

Farmer, T. a, Christiansen, M. H., & Monaghan, P. (2006). Phonological typicality influences on-

line sentence comprehension. Proceedings of the National Academy of Sciences of the

United States of America, 103, 12203–12208. doi:10.1073/pnas.0602173103

Fischer-Jørgensen, E. (1968). Perceptual dimensions of vowels. STUF-Language Typology and

Universals, 21, 94–98.

Fischer-Jørgensen, E. (1978). On the universal character of phonetic symbolism with special

184

reference to vowels. Studia Linguistica, 32, 80–90.

Fitch, W. T. (1997). Vocal tract length and formant frequency dispersion correlate with body size

in rhesus macaques. The Journal of the Acoustical Society of America, 102, 1213–1222.

Fitch, W. T., & Reby, D. (2001). The descended larynx is not uniquely human. Proceedings of

the Royal Society B, 268, 1669–1675.

Fitneva, S. A., Christiansen, M. H., & Monaghan, P. (2009). From sound to syntax: Phonological

constraints on children's lexical categorization of new words. Journal of Child

Language, 36, 967–997.

French, P. L. (1977). Towards an explanation of phonetic symbolism. Word, 28, 305–322.

doi:10.1080/00437956.1977.11435647

Fort, M., Lammertink, I., Peperkamp, S., Guevara‐Rukoz, A., Fikkert, P., & Tsuji, S. (2018).

Symbouki: a meta‐analysis on the emergence of sound symbolism in early language

acquisition. Developmental Science, 21, e12659.

Fort, M., Martin, A., & Peperkamp, S. (2015). Consonants are more important than vowels in the

bouba-kiki effect. Language and Speech, 58, 247–266.

Fort, M., Weiß, A., Martin, A., & Peperkamp, S. (2013, June). Looking for the bouba-kiki effect

in pre-lexical infants. Poster presented at the International Child Phonology Conference,

Nijmegen, the Netherlands. von der Gabelentz, G. (1891). Die sprachwissenschaft: Ihre aufgaben, methoden und bisherigen

ergebnisse. Leipzig: T.O. Weigel.

Gallace, A., Boschin, E., & Spence, C. (2011). On the taste of “Bouba” and “Kiki”: An

exploration of word-food associations in neurologically normal participants. Cognitive

Neuroscience, 2, 34–46. doi:10.1080/17588928.2010.516820

185

Gallace, A., & Spence, C. (2006). Multisensory synesthetic interactions in the speeded

classification of visual size. Perception & Psychophysics, 68, 1191–1203.

Galton, F. (1884). Measurement of character. Fortnightly Review, 36, 179–185.

Gamkrelidze, T. V. (1974). The problem of ‘l’arbitraire du signe’. Language, 50, 102–110.

Gelman, A., Jakulin, A., Pittau, M. G., & Su, Y. S. (2008). A weakly informative default prior

distribution for logistic and other regression models. The Annals of Applied Statistics, 2,

1360–1383.

Gelman, A., & Rubin, D. B. (1992). Inference from iterative simulation using multiple

sequences. Statistical Science, 7, 457–472.

Gasser, M. (2004). The origins of arbitrariness in language. Proceedings of the 26th Annual

Conference of the Cognitive Science Society, 4–7. Retrieved from http://www.cs.indiana.

edu/l/www/pub/gasser/cogsci04.pdf

Gentilucci, M., & Campione, G. C. (2011). Do postures of distal effectors affect the control of

actions of other distal effectors? Evidence for a system of interactions between hand and

mouth. PLoS ONE, 6. doi: 10.1371/journal.pone.0019793

Gentilucci M., & Corballis M. C. (2006). From manual gesture to speech: A gradual transition.

Neuroscience and Biobehavioral Reviews, 30, 949–960.

Gingras, B., Boeckle, M., Herbst, C. T., & Fitch, W. T. (2013). Call acoustics reflect body size

across four clades of anurans. Journal of Zoology, 289, 143–150.

Gold, R., & Segal, O. (2017). The bouba-kiki effect and its relation to the Autism Quotient (AQ)

in autistic adolescents. Research in Developmental Disabilities, 71, 11–17.

Goldberg, L. R. (1990). An alternative" description of personality": the big-five factor

structure. Journal of Personality and Social Psychology, 59, 1216–1229.

186

Gordon, M., & Heath, J. (1998). Sex, sound Symbolism, and sociolinguistics. Current

Anthropology, 39, 421–449.

Green, P. and MacLeod, C. J. (2016). simr: an R package for power analysis of generalised linear

mixed models by simulation. Methods in Ecology and Evolution, 7, 493–498.

doi:10.1111/2041-210X.12504

Greenberg, J. H. (1978). Introduction. In J. H. Greenberg, C. A. Ferguson, & E. A. Moravcsik

(Eds.), Universals of language, Volume 2: Phonology (pp. 1–8). Redwood City, CA:

Stanford University Press.

Greenberg, J. H., & Jenkins, J. J. (1966). Studies in the psychological correlates of the sound

system of American English. Word, 22, 207–242.

Hargreaves, I. S., & Pexman, P. M. (2012). Does richness lose its luster? Effects of extensive

practice on semantic richness in visual word recognition. Frontiers in Human

Neuroscience, 6, 234. doi: 10.3389/fnhum.2012.00234

Harm, M. W., & Seidenberg, M. S. (2004). Computing the meanings of words in reading:

cooperative division of labor between visual and phonological processes. Psychological

Review, 111, 662–720.

Hashimoto, T., Usui, N., Taira, M., Nose, I., Haji, T., & Kojima, S. (2006). The neural

mechanism associated with the processing of onomatopoeic sounds. Neuroimage, 31,

1762–1770.

Hauser, M. D. (1993). The evolution of nonhuman primate vocalizations: effects of phylogeny,

body weight, and social context. The American Naturalist, 142, 528–542.

Hino, Y., & Lupker, S. J. (1996). Effects of polysemy in lexical decision and naming: An

alternative to lexical access accounts. Journal of Experimental Psychology: Human

187

Perception and Performance, 22, 1331–1356.

Hinton, L., Nichols, J., & Ohala, J. J. (1994). Sound-symbolic processes. In L. Hinton, J. Nichols,

& J. Ohala (Eds.), Sound symbolism (pp. 1–14). New York, NY: Cambridge University

Press.

Hockett, C. (1963). The problem of universals in language. In J. Greenberg (Ed.), Universals of

language (pp. 1–22). Cambridge, MA: M.I.T. Press.

Hoffman, M. D., & Gelman, A. (2014). The No-U-turn sampler: adaptively setting path lengths

in Hamiltonian Monte Carlo. Journal of Machine Learning Research, 15, 1593–1623.

Hung, S. -M., Styles, S. J., & Hsieh, P. -J. (2017). Can a word sound like a sharp before you have

seen it? Sound-shape mapping prior to conscious awareness. Psychological Science, 28,

263–275 doi:10.1177/0956797616677313.

Hutton, C. (1989). The arbitrary nature of the sign. Semiotica, 75, 63–78.

Ikegami, T., & Zlatev, J. (2007). From non-representational cognition to language. In T. Ziemke,

J. Zlatev, & R. M. Frank (Eds.), Body, language and mind, vol 1: Embodiment (241–283).

Berlin: Mouton.

Imai, M., & Kita, S. (2014). The sound symbolism bootstrapping hypothesis for language

acquisition and language evolution. Philosophical Transactions of the Royal Society of

London. Series B, Biological Sciences, 369, 20130298.

Jakobson, R. (1965). Quest for the essence of language. Diogenes, 13, 21–37.

Jakobson, R., Fant, G., & Halle, M. (1951). Preliminaries to speech analysis. The distinctive

features and their correlates. Cambridge, MA: MIT Press.

Jakobson, R., & Waugh, L. (1979). The sound shape of language. Bloomington, IN: Indiana

University Press.

188

Jesperson, O. (1922). The symbolic value of the vowel i. Philologica, 1, 1–19.

Johansson, N., & Zlatev, J. (2013). Motivations for sound symbolism in spatial deixis: A

typological study of 101 languages. The Public Journal of Semiotics, 5, 3–20.

Johnson, R. C., Suzuki, N. S., & Olds, W. K. (1964). Phonetic symbolism in an artificial

language. The Journal of Abnormal and Social Psychology, 69, 233–236.

Joseph, J. E. (2015). Iconicity in Saussure's Linguistic Work, and why it does not contradict the

arbitrariness of the sign. Historiographia Linguistica, 42, 85–105.

Juhasz, B. J., & Yap, M. J. (2013). Sensory experience ratings for over 5,000 mono-and

disyllabic words. Behavior Research Methods, 45, 160–168. doi:10.3758/s13428-012-

0242-9

Juhasz, B. J., Yap, M. J., Dicke, J., Taylor, S. C., & Gullick, M. M. (2011). Tangible words are

recognized faster: The grounding of meaning in sensory and perceptual systems. The

Quarterly Journal of Experimental Psychology, 64, 1683–1691.

doi:10.1080/17470218.2011.605150

Kanero, J., Imai, M., Okuda, J., Okada, H., & Matsuda, T. (2014). How sound symbolism is

processed in the brain: A study on Japanese mimetic words. PLoS ONE, 9, 1–8.

Kantartzis, K. F. (2011). Children and adults’ understanding and use of sound-symbolism in

novel words (Doctoral dissertation). Retrieved from eTheses Repository. (2997)

Karwoski, T. F., Odbert, H. S., & Osgood, C. E. (1942). Studies in synesthetic thinking: II. The

role of form in visual responses to music. Journal of General Psychology, 26, 199–222.

Kawahara, S. & Kazuko S. (2012) A tripartite trans-module relationship between sounds, shapes

and emotions: A case of abrupt modulation. In Proceedings of the 34th Annual Meeting of

the Cognitive Science Society, Naomi Miyake, David Peebles, & Richard P. Cooper, eds.,

189

Austin, TX: Cognitive Science Society, 569–574.

Kawahara, S., Noto, A., & Kumagai, G. (2018). Sound symbolic patterns in Pokemon

names. Phonetica. doi:10.1159/000484938

Kawahara, S. & Kazuko S. (2012) A tripartite trans-module relationship between sounds, shapes

and emotions: A case of abrupt modulation. In Proceedings of the 34th Annual Meeting of

the Cognitive Science Society, Naomi Miyake, David Peebles, & Richard P. Cooper, eds.,

Austin, TX: Cognitive Science Society, 569–574.

Kawahara, S., Shinohara, K., & Grady, J. (2015). Iconic inferences about personality: From

sounds and shapes. In M. Hiraga, W. Herlofsky, K. Shinohara, & K. Akita (Eds.),

Icnonicity: East Meets West (pp. 57–70). John Benjamins Publishing Co.: Amsterdam.

Kingston, J., & Diehl, R. L. (1994). Phonetic knowledge. Language, 419–454.

Kirkham, N. Z., Slemner, J. A., & Johnson, S. P. (2002). Visual statistical learning in infancy:

Evidence for a domain general learning mechanism. Cognition, 83, B35–B42.

Klank, L. J., Huang, Y. H., & Johnson, R. C. (1971). Determinants of success in matching word

pairs in tests of phonetic symbolism. Journal of Verbal Learning and Verbal

Behavior, 10, 140–148.

Klink, R. R. (2000). Creating brand names with meaning: The use of sound

symbolism. Marketing Letters, 11, 5–20.

Klink, R. R., & Athaide, G. A. (2012). Creating brand personality with brand names. Marketing

Letters, 23, 109–117.

Köhler, W. (1929). Gestalt Psychology. New York, USA: Liveright.

Komeilipoor, N., Tiainen, M., Tiippana, K., Vainio, M., & Vainio, L. (2016). Excitability of hand

motor areas during articulation of syllables. Neuroscience Letters, 620, 154–158.

190

Koriat, A., & Levy, I. (1977). The symbolic implications of vowels and of their orthographic

representations in two natural languages. Journal of Psycholinguistic Research, 6, 93–

103.

Kovic, V., Plunkett, K., & Westermann, G. (2010). The shape of words in the brain. Cognition,

114, 19–28.

Kruschke, J. K. (2015). Doing bayesian data analysis: A tutorial with R, JAGS, and stan (2nd.

ed.). Boston: Academic Press.

Kuperman, V., Stadthagen-Gonzalez, H., & Brysbaert, M. (2012). Age-of-acquisition ratings for

30,000 English words. Behavior Research Methods, 44, 978–990.

Kuznetsova, A., Brockhoff, P.B., & Christensen, R.H.B. (2017). “lmerTest Package: Tests in

Linear Mixed Effects Models.” Journal of Statistical Software, 82, 1–26.

Ladefoged, P., & Johnson, K. (2010). A Course in Linguistics (6 ed.). Boston MA: Wadsworth.

Laham, S. M., Koval, P., & Alter, A. L. (2012). The name-pronunciation effect: Why people like

Mr. Smith more than Mr. Colquhoun. Journal of Experimental Social Psychology, 48,

752–756.

Laing, C.E., Viham, M. & Portnoy, T.K. (2017). How salient are onomatopoeia in the early

input? A prosodic analysis of infant-directed speech. Journal of Child Language, 44,

1117–1139.

Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato's problem: The latent semantic

analysis theory of acquisition, induction, and representation of knowledge. Psychological

Review, 104, 211–240.

Lee, K., & Ashton, M. C. (2004). Psychometric properties of the HEXACO personality

inventory. Multivariate Behavioral Research, 39, 329–358.

191

Lee, K., & Ashton, M. C. (2006). Further assessment of the HEXACO Personality Inventory:

Two new facet scales and an observer report form. Psychological Assessment, 18, 182–

191.

Lee, K., & Ashton, M. C. (2008). The HEXACO personality factors in the indigenous personality

lexicons of English and 11 other languages. Journal of Personality, 76, 1001–1054.

Lee, K., & Ashton, M. C. (2009). Scale descriptions. Retrieved from

http://hexaco.org/scaledescriptions

Lee, K., & Ashton, M. C. (2016). Psychometric properties of the HEXACO-100. Assessment.

doi: 10.1177/1073191116659134

Lee, K., & Ashton, M. C. (2017). Acquaintanceship and self/observer agreement in personality

judgment. Journal of Research in Personality, 70, 1–5.

Leirer, V. O., Hamilton, D. L., & Carpenter, S. (1982) Common first names as cues for

inferences about personality. Personality and Social Psychology Bulletin, 8, 712–718.

Lewkowicz, D. J., & Turkewitz, G. (1980). Cross-modal equivalence in early infancy: Auditory–

visual intensity matching. Developmental Psychology, 16, 597–607.

Lockwood, G., & Dingemanse, M. (2015). Iconicity in the lab: A review of behavioral,

developmental, and neuroimaging research into sound-symbolism. Frontiers in

Psychology, 6. doi:10.3389/fpsyg.2015.01246

Lockwood, G., Dingemanse, M., & Hagoort, P. (2016). Sound-symbolism boosts novel word

learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 42,

1274.

Lockwood, G., Hagoort, P., & Dingemanse, M. (2016). How Iconicity Helps People Learn New

Words : Neural Correlates and Individual Differences in Sound-Symbolic Bootstrapping,

192

Collabra, 2, 1–15.

Lockwood, G., & Tuomainen, J. (2015). Ideophones in Japanese modulate the P2 and late

positive complex responses. Frontiers in Psychology, 6. doi:10.3389/fpsyg.2015.00933

Lowrey, T. M., & Shrum, L. J. (2007). Phonetic symbolism and brand name preference. Journal

of Consumer Research, 34, 406–414.

Ludwig, V. U., Adachi, I., & Matsuzawa, T. (2011). Visuoauditory mappings between high

luminance and high pitch are shared by chimpanzees (Pan troglodytes) and

humans. Proceedings of the National Academy of Sciences, 108, 20661-20665.

Lupyan, G., & Casasanto, D. (2015). Meaningless words promote meaningful categorization.

Language and Cognition, 7, 167–193.

Lupyan, G., & Winter, B. (2018). Language is more abstract than you think, or, why aren't

languages more iconic?. Philosophical Transactions of the Royal Society B: Biological

Sciences, 373, 20170137.

Lynott, D., & Connell, L. (2009). Modality exclusivity norms for 423 object properties. Behavior

Research Methods, 41, 558–564. doi:10.3758/BRM.41.2.558

Lynott, D., & Connell, L. (2013). Modality exclusivity norms for 400 nouns: The relationship

between perceptual experience and surface word form. Behavior Research Methods, 45,

516–526.

MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual Review of

Psychology, 58, 593–614.

Maeda, F., Kanai, R., & Shimojo, S. (2004). Changing pitch induced visual motion illusion.

Current Biology, 14, R990–R991.

193

Maglio, S. J., Rabaglia, C. D., Feder, M. A., Krehm, M., & Trope, Y. (2014). Vowel sounds in

words affect mental construal and shift preferences for targets. Journal of Experimental

Psychology: General, 143, 1082–1096.

Magnus, M. (2000). What's in a Word? Evidence for phonosemantics. (Doctoral dissertation).

Retrieved from NTNU Open. (82-471-5073-5)

Marks, L. E. (1974). On associations of light and sound: The mediation of brightness, pitch, and

loudness. The American Journal of Psychology, 173–188.

Marks, L. E. (1987). On cross-modal similarity: Auditory–visual interactions in speeded

discrimination. Journal of Experimental Psychology: Human Perception and

Performance, 13, 384–394.

Marks, L. E. (1989). On cross-modal similarity: the perceptual structure of pitch, loudness, and

brightness. Journal of Experimental Psychology: Human Perception and

Performance, 15, 586–602.

Martino, G., & Marks, L. E. (1999). Perceptual and linguistic interactions in speeded

classification: Tests of the semantic coding hypothesis. Perception, 28, 903–923.

Masuda, K. (2007). The physical basis for phonological iconicity. In E. Tabakowska, C.

Ljungberg, & O. Fischer (Eds.), Insistent Images (pp. 57–72). Philadephia, PA: John

Benjamins.

Maurer, D., Pathman, T., & Mondloch, C. J. (2006). The shape of boubas: Sound-shape

correspondences in toddlers and adults. Developmental Science, 9, 316–322.

doi:10.1111/j.1467-7687.2006.00495.x

McCormick, K., Kim, J. Y., List, S., & Nygaard, L. C. (2015). Sound to meaning mappings in the

bouba–kiki effect. In D. C. Noelle, R. Dale, A. S. Warlaumont, J. Yoshimi, T. Matlock,

194

C. D. Jennings, & P. P. Maglio (Eds.), Proceedings of the 37th annual conference of the

cognitive science society (pp. 1565–1570). Austin, TX: Cognitive Science Society.

Mehrabian, A., & Piercy, M. (1993a). Affective and personality characteristics inferred from

length of first names. Personality and Social Psychology Bulletin, 19, 755–758.

Mehrabian, A., & Piercy, M. (1993b). Positive or negative connotations of unconventionally and

conventionally spelled names. The Journal of Social Psychology, 133, 445–451.

Meir, I., Padden, C., Aronoff, M., & Sandler, W. (2013). Competing iconicities in the structure of

languages. Cognitive Linguistics, 24, 309–343.

Meteyard, L., Stoppard, E., Snudden, D., Cappa, S. F., & Vigliocco, G. (2015). When semantics

aids phonology: A processing advantage for iconic word forms in

aphasia. Neuropsychologia, 76, 264–275.

Miron, M. S. (1961). A crosslinguistic investigation of phonetic symbolism. The Journal of

Abnormal and Social Psychology, 62, 623–630. doi:10.1037/h0045212

Misyak, J. B., & Christiansen, M. H. (2012). Statistical learning and language: An individual

differences study. Language Learning, 62, 302–331.

Monaghan, P., Christiansen, M. H., & Fitneva, S. A. (2011). The arbitrariness of the sign:

learning advantages from the structure of the vocabulary. Journal of Experimental

Psychology: General, 140, 325–347.

Monaghan, P., & Fletcher, M. (2019). Do sound symbolism effects for written words relate to

individual phonemes or to phoneme features?. Language and Cognition, 11, 235–255.

Monaghan, P., Mattock, K., & Walker, P. (2012). The role of sound symbolism in language

learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38,

1152–1164.

195

Monaghan, P., Shillcock, R. C., Christiansen, M. H., & Kirby, S. (2014). How arbitrary is

language?. Philosophical Transactions of the Royal Society of London. Series B,

Biological Sciences, 369, 20130299. doi:10.1098/rstb.2013.0299

Moos, A., Smith, R., Miller, S. R., & Simmons, D. R. (2014). Cross-modal associations in

synaesthesia: vowel colours in the ear of the beholder. i-Perception, 5, 132–142.

Morton, E. S. (1977). On the occurrence and significance of motivation-structural rules in some

bird and mammal sounds. American Naturalist, 855–869.

Morton, E. S. (1994). Sound symbolism and its role in non-human vertebrate communication. In

L. Hinton, J. Nichols, & J. Ohala (Eds.), Sound Symbolism (pp. 348–365). New York,

NY: Cambridge University Press.

Newman, S. S. (1933). Further experiments in phonetic symbolism. The American Journal of

Psychology, 45, 53–75.

Nichols, J. (1971). Diminutive consonant symbolism in western North America. Language, 47,

826–4.

Nielsen, A. K. S., & Rendall, D. (2011). The sound of round: evaluating the sound-symbolic role

of consonants in the classic Takete-Maluma phenomenon. Canadian Journal of

Experimental Psychology, 65, 115–124. doi:10.1037/a0022268

Nielsen, A. K. S., & Rendall, D. (2013). Parsing the role of consonants versus vowels in the

classic Takete-Maluma phenomenon. Canadian Journal of Experimental Psychology, 67,

153–163. doi:10.1037/a0030553

Nygaard, L. C., Cook, A. E., & Namy, L. L. (2009). Sound to meaning correspondences facilitate

word learning. Cognition, 112, 181–186. doi:10.1016/j.cognition.2009.04.001

Oberman, L. M., & Ramachandran, V. S. (2008). Preliminary evidence for deficits in

196

multisensory integration in autism spectrum disorders: The mirror neuron

hypothesis. Social Neuroscience, 3, 348–355.

Occelli, V., Esposito, G., Venuti, P., Arduino, G. M., & Zampini, M. (2013). The Takete—

Maluma Phenomenon in Autism Spectrum Disorders. Perception, 42, 233–241.

Oh, D., Buck, E. A., & Todorov, A. (2019). Revealing Hidden Gender Biases in Competence

Impressions of Faces. Psychological Science, 30, 65–79.

Ohala, J. J. (1994). The frequency code underlies the sound-symbolic use of voice pitch. In L.

Hinton, J. Nichols, & J. Ohala (Eds.), Sound Symbolism (pp. 325–347). New York, NY:

Cambridge University Press.

Ohala, J. J., & Eukel, B. W. (1987). Explaining the intrinsic pitch of vowels. In R. Channon & L.

Shockey (Eds.). In honour of Ilse Lehiste (pp. 207–215). Dordrecht: Foris.

Ohtake, Y., & Haryu, E. (2013). Investigation of the process underpinning vowel-size

correspondence. Japanese Psychological Research, 55, 390–399. doi:10.1111/jpr.12029

Osgood, C.E., Suci, S. J., & Tannenbaum, P. H. (1957). The Measurement of Meaning. Urbana,

IL: Univ. of Illinois Press.

Osgood, C. E. (1963). Language universals and psycholinguistics. In J. Greenberg (Ed.),

Universals of Language (pp. 299–328). Cambridge, Mass.: M.I.T. Press.

Owren, M. J., & Rendall, D. (2001). Sound on the rebound: bringing form and function back to

the forefront in understanding nonhuman primate vocal signaling. Evolutionary

Anthropology: Issues, News, and Reviews, 10, 58–71.

Ozturk, O., Krehm, M., & Vouloumanos, A. (2013). Sound symbolism in infancy: evidence for

sound-shape cross-modal correspondences in 4-month-olds. Journal of Experimental

197

Child Psychology, 114, 173–186.

Parise, C. V. (2016). Crossmodal correspondences: Standing issues and experimental

guidelines. Multisensory Research, 29, 7–28.

Parise, C. V., & Spence, C. (2012). Audiovisual crossmodal correspondences and sound

symbolism: A study using the implicit association test. Experimental Brain Research,

220, 319–333. doi:10.1007/s00221-012-3140-6

Parise, C. V., & Spence, C. (2013). Audiovisual cross-modal correspondences in the general

population. In J. Simner, & E. Hubbard (Eds.), Oxford Handbook of Synaesthesia (pp.

790–815). Oxford, UK: Oxford University Press.

Patel, R., Mulder, R. A., & Cardoso, G. C. (2010). What makes vocalisation frequency an

unreliable signal of body size in birds? A study on black swans. Ethology, 116, 554–563.

Peeters, D. (2016). Processing consequences of onomatopoeic iconicity in spoken language

comprehension. In 38th Annual Meeting of the Cognitive Science Society (CogSci

2016) (pp. 1632–1647). Cognitive Science Society.

Peirce, C. S. (1974). Collected papers of Charles Sanders Peirce (6 ed.). Boston MA: Harvard

University Press.

Pejovic, J., & Molnar, M. (2016). The Development of Spontaneous Sound-shape Matching in

Monolingual and Bilingual Infants During the First Year. Developmental Psychology.

doi:10.1037/dev0000237

Peña, M., Mehler, J., & Nespor, M. (2011). The role of audiovisual processing in early

conceptual development. Psychological Science, 22, 1419–1421.

Perlman, M., Dale, R., & Lupyan, G. (2015). Iconicity can ground the creation of vocal

symbols. Royal Society Open Science, 2, 150152.

198

Perlman, M., & Lupyan, G. (2018). People can create iconic vocalizations to communicate

various meanings to naïve listeners. Scientific Reports, 8, 2634.

Perry, L. K., Perlman, M., & Lupyan, G. (2015). Iconicity in english and Spanish and its relation

to lexical category and age of acquisition. PLoS ONE, 10.

doi:10.1371/journal.pone.0137147

Perry, L. K., Perlman, M., Winter, B., Massaro, D. W., & Lupyan, G. (2017). Iconicity in the

speech of children and adults. Developmental Science, 21. doi:10.1111/desc.12572

Perniss, P., Lu, J. C., Morgan, G., & Vigliocco, G. (2018). Mapping language to the world: the

role of iconicity in the sign language input. Developmental Science, 21. doi:

10.1111/desc.12551

Perniss, P., Thompson, R. L., & Vigliocco, G. (2010). Iconicity as a general property of

language: Evidence from spoken and signed languages. Frontiers in Psychology, 1.

doi:10.3389/fpsyg.2010.00227

Perniss, P., & Vigliocco, G. (2014). The bridge of iconicity: from a world of experience to the

experience of language. Philosophical Transactions of the Royal Society of London.

Series B, Biological Sciences, 369, 20130300.

Pexman, P. M., Heard, A., Lloyd, E., & Yap, M. J. (2017). The Calgary semantic decision

project: concrete/abstract decision data for 10,000 English words. Behavior Research

Methods, 49, 407–417.

Pexman, P. M., Lupker, S. J., & Jared, D. (2001). Homophone effects in lexical decision. Journal

of Experimental Psychology: Learning, Memory, & Cognition, 27, 139–156.

Pexman, P. M., Lupker, S. J., & Reggin, L. D. (2002). Phonological effects in visual word

recognition: Investigating the impact of feedback activation. Journal of Experimental

199

Psychology: Learning, Memory, & Cognition, 28, 572–584.

Pexman, P. M., Siakaluk, P. D., & Yap, M. J. (2014). Introduction to the research topic Meaning

in Mind: Semantic richness effects in language processing. In P. M. Pexman, P. D.

Siakaluk, & M. J. Yap (Eds). Meaning in Mind: Semantic Richness Effects in Language

Processing. Lausanne, Switzerland: Frontiers E-Books. doi: 10.3389/fnhum.2013.00723.

Preziosi, M. A., & Coane, J. H. (2017). Remembering that big things sound big: Sound

symbolism and associative memory. Cognitive Research: Principles and Implications, 2.

doi:10.1186/s41235-016-0047-y

R Core Team (2018). R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.

Rabaglia, C. D., Maglio, S. J., Krehm, M., Seok, J. H., & Trope, Y. (2016). The sound of

distance. Cognition, 152, 141–149. doi:10.1016/j.cognition.2016.04.001

Ramachandran, V. S., & Hubbard, E. M. (2001). Synaesthesia: A window into perception,

thought and language. Journal of Consciousness Studies, 8, 3–34.

Ramachandran, V. S., & Hubbard, E. M. (2005). The emergence of the human mind: Some clues

from synesthesia. In L. C. Robertson & N. Sagiv (Eds.), Synesthesia: Perspectives from

cognitive neuroscience (pp. 147–192). Oxford: Oxford University Press.

Reichard, G. A. (1944). Prayer: The compulsive word. American Ethnological Society

Monograph, 7. Seattle, WA: University of Washington Press.

Reichard, G. A. (1950). Navaho religion: A study of symbolism. New York, NY: Pantheon

Books.

Reilly, J., & Kean, J. (2007). Formal distinctiveness of high-and low-imageability nouns:

Analyses and theoretical implications. Cognitive Science, 31, 157–168.

200

Rendall, D. (2003). Acoustic correlates of caller identity and affect intensity in the vowel-like

grunt vocalizations of baboons. The Journal of the Acoustical Society of America, 113,

3390–3402.

Rendall, D., Kollias, S., Ney, C., & Lloyd, P. (2005). Pitch (F0) and formant profiles of human

vowels and vowel-like baboon grunts: the role of vocalizer body size and voice-acoustic

allometry. The Journal of the Acoustical Society of America, 117, 944–955.

Reetz, H., & Jongman, A. (2009). Phonetics: Transcription, production, acoustics, and

perception. Hoboken, NJ: WileyBlackwell.

Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits

from childhood to old age: a quantitative review of longitudinal studies. Psychological

bulletin, 126, 3–25.

Rogers, S. K., & Ross, A. S. (1975). A cross-cultural test of the Maluma-Takete phenomenon.

Perception, 4, 105–106.

Rummer, R., Schweppe, J., Schlegelmilch, R., & Grice, M. (2014). Mood is linked to vowel type:

the role of articulatory movements. Emotion, 14, 246–250. doi:10.1037/a0035752

Saji, N., Akita, K., Imai, M., Kantartzis, K., & Kita, S. (2013). Cross-linguistically shared and

language-specific sound symbolism for motion: an exploratory data mining approach.

Proc. 35th Annual Conf. of the Cognitive Science Society, Berlin, Germany, 31, 1253–

1259.

Sapir, E. (1929). A study in phonetic symbolism. Journal of Experimental Psychology, 12, 225–

239.

Sedley, D. (2006). Plato’s Cratylus. In E. N. Zalta (Ed.), The Stanford encyclopedia of philosophy

(Fall 2013 ed.). Retrieved from https://plato.stanford.edu/entries/plato-cratylus/

201

Sereno, J. A. (1986). Stress pattern differentiation of form class in English. The Journal of the

Acoustical Society of America, 79, S36–S36.

Shaoul, C., & Westbury, C. (2010). Exploring lexical co-occurrence space using HiDEx. Behavior

Research Methods, 42, 393–413. doi:10.3758/Brm.42.2.393

Shayan, S., Ozturk, O., & Sicoli, M. A. (2011). The thickness of pitch: Crossmodal metaphors in

Farsi, Turkish, and Zapotec. The Senses and Society, 6, 96–105.

Shinohara, K., & Kawahara, S. (2016). A cross-linguistic study of sound symbolism: The images

of size. Proceedings of the 36th Annual Meeting of the Berkeley Linguistics Society.

Shintel, H., Nusbaum, H. C., & Okrent, A. (2006). Analog acoustic expression in speech

communication. Journal of Memory and Language, 55, 167–177.

Sidhu, D. M., & Pexman, P. M. (2015). What's in a name? Sound symbolism and gender in first

names. PloS ONE, 10. doi:10.1371/journal.pone.0126809

Sidhu, D. M., & Pexman, P. M. (2016). A Prime Example of the Maluma/Takete Effect? Testing

for Sound Symbolic Priming. Cognitive Science. doi:10.1111/cogs.12438

Sidhu, D. M., & Pexman, P. M. (2018a). Five mechanisms of sound symbolic

association. Psychonomic Bulletin & Review, 25, 1619–1643.

Sidhu, D. M., & Pexman, P. M. (2018b). Lonely sensational icons: semantic neighbourhood

density, sensory experience and iconicity. Language, Cognition and Neuroscience, 33,

25–31.

Sidhu, D. M., Pexman, P. M., & Saint-Aubin, J. (2016). From the Bob/Kirk effect to the

Benoit/Éric effect: Testing the mechanism of name sound symbolism in two

languages. Acta Psychologica, 169, 88–99.

Singmann, H., Bolker, B., & Westfall, J. (2015). afex: Analysis of factorial experiments. R

202

package version 0.23–0.

Slepian, M. L., & Galinsky, A. D. (2016). The voiced pronunciation of initial phonemes predicts

the gender of names. Journal of Personality and Social Psychology, 110, 509–527.

Smith, R. (2004). Fitting Sense to Sound: Linguistic Aesthetics and Phonosemantics in the Work

of JRR Tolkien. Tolkien Studies, 3, 1–20.

Smith, L. B., & Sera, M. D. (1992). A developmental analysis of the polar structure of

dimensions. Cognitive Psychology, 24, 99–142.

Spence, C. (2011). Crossmodal correspondences: A tutorial review. Attention, Perception, &

Psychophysics, 73, 971–995. doi:10.3758/s13414-010-0073-7

Spence, C. (2012). Managing sensory expectations concerning products and brands: Capitalizing

on the potential of sound and shape symbolism. Journal of Consumer Psychology, 22, 37–

54.

Spence, C., & Gallace, A. (2011). Tasting shapes and words. Food Quality and Preference, 22,

290–295.

Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human

smile: a nonobtrusive test of the facial feedback hypothesis. Journal of Personality and

Social Psychology, 54, 768–777.

Stevens, S. S. (1957). On the psychophysical law. Psychological Review, 64, 153–181.

Styles, S. J., & Gawne, L. (2017). When does maluma/takete fail? Two key failures and a meta-

analysis suggest that phonology and phonotactics matter. i-Perception, 8.

doi:10.1177/2041669517724807

Sučević, J., Savić, A. M., Popović, M. B., Styles, S. J., & Ković, V. (2015). Balloons and

bavoons versus spikes and shikes: ERPs reveal shared neural processes for shape-sound-

203

meaning congruence in words, and shape-sound congruence in pseudowords. Brain and

Language, 145/146, 11–22. doi:10.1016/j.bandl.2015.03.011

Tanz, C. (1971). Sound symbolism in words relating to proximity and distance. Language and

Speech, 14, 266–276.

Tarte, R. D. (1982). The relationship between monosyllables and pure tones: An investigation of

phonetic symbolism. Journal of Verbal Learning and Verbal Behavior, 21, 352–360.

doi:10.1016/S0022-5371(82)90670-3

Taub, S. F. (2001). Language from the body: Iconicity and metaphor in American Sign

Language. Cambridge, UK: Cambridge University Press.

Taylor, I. K. (1963). Phonetic symbolism revisited. Psychological Bulletin, 60, 200–209.

Teramoto, W., Hidaka, S., & Sugita, Y. (2010). Sounds move a static visual object. PLoS

ONE, 5. doi: 10.1371/journal.pone.0012255

Thompson, P. D., & Estes, Z. (2011). Sound symbolic naming of novel objects is a graded

function. The Quarterly Journal of Experimental Psychology, 64, 37–41.

doi:10.1080/17470218.2011.605898

Topolinski, S., Maschmann, I. T., Pecher, D., & Winkielman, P. (2014). Oral approach–

avoidance: Affective consequences of muscular articulation dynamics. Journal of

Personality and Social Psychology, 106, 885–896.

Tsur, R. (2006). Size-sound symbolism revisited. Journal of Pragmatics, 38, 905–924.

doi:10.1016/j.pragma.2005.12.002

Tucker, M., & Ellis, R. (2001). The potentiation of grasp types during visual object

categorization. Visual Cognition, 8, 769–800.

204

Tupes, E. C., & Christal, R. E. (1992). Recurrent personality factors based on trait

ratings. Journal of Personality, 60, 225–251.

Tzeng, C. Y., Nygaard, L. C., & Namy, L. L. (2017). The Specificity of Sound Symbolic

Correspondences in Spoken Language. Cognitive Science, 41, 2191–2220.

Tyler, M. D., Best, C. T., Faber, A., & Levitt, A. G. (2014). Perceptual assimilation and

discrimination of non-native vowel contrasts. Phonetica, 71, 4–21.

Urban, M. (2011). Conventional sound symbolism in terms for organs of speech: A cross-

linguistic study. Folia Linguist, 45, 199–214.

Vainio, L., Schulman, M., Tiippana, K., & Vainio, M. (2013). Effect of Syllable Articulation on

Precision and Power Grip Performance. PLoS ONE, 8. doi:10.1371/journal.pone.0053061

Vainio, L., Tiainen, M., Tiippana, K., Rantala, A., & Vainio, M. (2016). Sharp and round shapes

of seen objects have distinct influences on vowel and consonant articulation.

Psychological Research. doi:10.1007/s00426-016-0778-x

Valentine, B., Brennen, T., & Brédart, S. (1996). The cognitive psychology of proper names: On

the importance of being Ernest. New York: Routledge.

Van Lancker, D., & Cummings, J. L. (1999). Expletives: Neurolinguistic and neurobehavioral

perspectives on swearing. Brain Research Reviews, 31, 83–104.

Velasco, C., Salgado-Montejo, A., Marmolejo-Ramos, F., & Spence, C. (2014). Predictive

packaging design: Tasting shapes, typefaces, names, and sounds. Food Quality and

Preference, 34, 88–95.

Verhoef, T., Roberts, S. G., & Dingemanse, M. (2015). Emergence of systematic iconicity:

transmission, interaction and analogy. Proceedings of the 37th Annual Meeting of the

Cognitive Science Society, 2481–2486.

205

Vernon, R. J., Sutherland, C. A., Young, A. W., & Hartley, T. (2014). Modeling first impressions

from highly variable facial images. Proceedings of the National Academy of

Sciences, 111, E3353–E3361.

Vigliocco, G., & Kita, S. (2006). Language-specific properties of the lexicon: Implications for

learning and processing. Language and Cognitive Processes, 21, 790–816.

Von Humboldt, W. (1836). On language: On the diversity of human language construction and

its influence on the mental development of the human species. Cambridge, UK:

Cambridge University Press.

Wagenmakers, E. J., Beek, T., Dijkhoff, L., Gronau, Q. F., Acosta, A., Adams, R. B., ... &

Bulnes, L. C. (2016). Registered Replication Report Strack, Martin, & Stepper

(1988). Perspectives on Psychological Science, 11, 917–928.

Walker, P. (2012). Cross-sensory correspondences and cross talk between dimensions of

connotative meaning: Visual angularity is hard, high-pitched, and bright. Attention,

Perception, & Psychophysics, 74, 1792–1809.

Walker, P. (2016). Cross-sensory correspondences and symbolism in spoken and written

language. Journal of Experimental Psychology: Learning, Memory, and Cognition, 42,

1339–1361.

Walker, P., & Walker, L. (2012). Size–brightness correspondence: Crosstalk and congruity

among dimensions of connotative meaning. Attention, Perception, & Psychophysics, 74,

1226–1240.

Walker, L., Walker, P., & Francis, B. (2012). A common scheme for cross-sensory

correspondences across stimulus domains. Perception, 41, 1186–1192.

206

Wallschläger, D. (1980). Correlation of song frequency and body weight in passerine

birds. Cellular and Molecular Life Sciences, 36, 412–412.

Walsh, V. (2003). A theory of magnitude: Common cortical metrices of time, space and quality.

Trends in Cognitive Sciences, 7, 483–488.

Warriner, A. B., Kuperman, V., & Brysbaert, M. (2013). Norms of valence, arousal, and

dominance for 13,915 English lemmas. Behavior Research Methods, 45, 1191–1207.

Wan, X., Woods, A. T., van den Bosch, J. J. F., McKenzie, K. J., Velasco, C., & Spence, C.

(2014). Cross-cultural differences in crossmodal correspondences between basic tastes

and visual features. Frontiers in Psychology, 5. doi:10.3389/fpsyg.2014.01365

Watkins, K. E., Strafella, A. P., & Paus, T. (2003). Seeing and hearing speech excites the motor

system involved in speech production. Neuropsychologia, 41, 989–994.

doi:10.1016/S0028-3932(02)00316-0

Waugh, L. R. (1992). Presidential address: let’s take the con out of iconicity. American Journal

of Semiotics, 9, 7–47.

Waugh, L. R. (1993). Against arbitrariness: Imitation and Motivation revived, with consequences

for textual meaning. Diacritics, 23, 71–87.

Weide, R. (2005). The Carnegie Mellon pronouncing dictionary.

Werner, H., & Kaplan, B. (1963). Symbol formation. An organismic-developmental approach to

language and the expression of thought. New York: Wiley.

Westbury, C. (2005). Implicit sound symbolism in lexical access: Evidence from an interference

task. Brain and Language, 93, 10–19. doi:10.1016/j.bandl.2004.07.006

Westbury, C., Hollis, G., Sidhu, D. M., & Pexman, P. M. (2018). Weighing up the evidence for

sound symbolism: Distributional properties predict cue strength. Journal of Memory and

207

Language, 99, 122–150.

Westermann, D. H. (1927). Laut, Ton und Sinn in westafrikanischen Sudansprachen. In F. Boas

(Ed.), Festschrift Meinhof. Hamburg: L. Friederichsen.

Wichmann, S., Holman, E. W., & Brown, C. H. (2013). Sound symbolism in basic vocabulary.

Entropy, 15, 844–858. doi:10.3390/e12040844

Winter, B. (2016). Taste and smell words form an affectively loaded and emotionally flexible

part of the English lexicon. Language, Cognition and Neuroscience, 31, 975–988.

Winter, B., Perlman, M., Perry, L. K., & Lupyan, G. (2017). Which words are most

iconic?. Interaction Studies, 18, 443–464.

Witherspoon, G. (1977). Language and art in the Navajo universe. Ann Arbor, MI: University of

Michigan Press.

Wu, L., Klink, R. R., & Guo, J. (2013). Creating gender brand personality with brand names: The

effects of phonetic symbolism. Journal of Marketing Theory and Practice, 21, 319–330.

Yap, M. J. (2007). Visual word recognition: Explorations of megastudies, multisyllabic words,

and individual differences. Unpublished doctoral dissertation, Washington University in

St. Louis, St. Louis, MO.

Yap, M. J., Lim, G. Y., & Pexman, P. M. (2015). Semantic richness effects in lexical decision:

the role of feedback. Memory & Cognition, 43, 1148–1167.

Yap, M. J., Pexman, P. M., Wellsby, M., Hargreaves, I. S., & Huff, M. J. (2012). An abundance

of riches: cross-task comparisons of semantic richness effects in visual word recognition.

Frontiers in Human Neuroscience, 6. doi:10.3389/fnhum.2012.00072

Yarkoni, T., Balota, D., & Yap, M. (2008). Moving beyond Coltheart’s N: A new measure of

orthographic similarity. Psychonomic Bulletin & Review, 15, 971–979.

208

Young, R. K., Kennedy, A. H., Newhouse, A., Browne, P., & Thiessen, D. (1993). The effects of

names on perception of intelligence, popularity, and competence. Journal of Applied

Social Psychology, 23, 1770–1788.

Zajonc, R. B., Murphy, S. T., & Inglehart, M. (1989). Feeling and facial efference: Implications

of the vascular theory of emotion. Psychological Review, 96, 395–416. doi:10.1037/0033-

295X.96.3.395

Zangenehpour, S., & Zatorre, R. J. (2010). Cross-modal recruitment of primary visual cortex

following brief exposure to bimodal audiovisual stimuli. Neuropsychologia, 48, 591–600.

209

Appendix A

Iconicity Taxonomy

Iconicity

Imagic Diagrammatic Relationship between one form and one meaning. Relationship between the relationship of forms and meanings.

B) C)

Direct Indirect Gestalt Relative Form directly resembles meaning. Form resembles meaning via an association. Word internal relationships. Between word relationships.

Arbitrary Non-Arbitrary No similarity in form and meaning relationships. Similarity in form and meaning relationships.

Figure A1. A proposed taxonomy of iconicity. In all cases of iconicity there is a structural resemblance-based mapping between form and meaning; however, this can happen in different ways. The above figure is an attempt to integrate the various distinctions that have been made in the literature. A) In imagic iconicity, the iconic relationship is between a single form and its meaning; in diagrammatic iconicity, the iconic relationship is between the relationships among forms and the relationships among their meanings (i.e., a relationship between relationships; Peirce, 1974). Note that this distinction is sometimes referred to as being between absolute and relative iconicity (e.g., Dingemanse et al., 2015). B) In direct imagic iconicity a form directly resembles its meaning (e.g.,

210 onomatopoeia); in indirect imagic iconicity a form resembles its meaning by way of its associations (e.g., words with sound symbolically small phonemes [e.g., /i/] with small meanings; Masuda, 2007).38 C) In gestalt diagrammatic iconicity, the form relationship is within a single word (e.g., word structure iconically conveying event structure; for instance, closed syllables evoking end points); in relative diagrammatic iconicity the form relationship is between two (or more) words (Dingemanse, 2012a). D)

Arbitrary relative iconicity refers to the mere existence of similar forms with similar meanings (e.g., phonesthemes); non-arbitrary relative iconicity refers to instances in which the relationship among word forms is similar to the relationship among their meanings

(e.g., three words with progressively higher vowels, meaning progressively smaller things).

38 While we attribute the term indirect iconicity–meaning sound symbolically mediated iconicity–to Masuda (2007), that author did not use it specifically in the context of imagic iconicity as we have here.

211

Appendix B

Stimuli from Chapter 3

Table B1

Trait stimuli used in Experiments 1, 2, and 4.

Trait Factor Direction Honest Honest-Humility High Sincere Honest-Humility High Trustworthy Honest-Humility High Conceited Honest-Humility Low Self-centered Honest-Humility Low Snobbish Honest-Humility Low Emotional Emotionality High Sensitive Emotionality High Sentimental Emotionality High Fearless Emotionality Low Rugged Emotionality Low Unemotional Emotionality Low Lively Extraversion High Outgoing Extraversion High Social Extraversion High Antisocial Extraversion Low Dull Extraversion Low Withdrawn Extraversion Low Agreeable Agreeableness High Cooperative Agreeableness High Peaceful Agreeableness High Aggressive Agreeableness Low Blunt Agreeableness Low Quick-tempered Agreeableness Low Hard-Working Conscientiousness High Organized Conscientiousness High Thorough Conscientiousness High Careless Conscientiousness Low Disorganized Conscientiousness Low Irresponsible Conscientiousness Low Complex Openness High Insightful Openness High Philosophical Openness High Conventional Openness Low Narrow-minded Openness Low Simple Openness Low

212

Table B2

Name stimuli used in Experiments 1 and 2, along with their frequencies (viz. number of baby names given in Alberta in 2014) and invented name transformations used in Experiment 4. Note that real names are shown in the pairs in which they were presented in Experiment 1.

Sonorant Invented Voiceless Stop Invented Name Frequency Name Name Frequency Name Abel 31 Aleb Eric 42 Erip Allen 2 Ammel Hector 3 Hepker Anne 4 Ull Rita 1 Reepa Joanna 18 Noaja Erica 7 Ekira June 10 Nuje Etta 5 Eppa Lanah 1 Namah Patty 1 Teeka Laurel 3 Maurem Christie 1 Triski Lauren 55 Mauren Katie 38 Tatie Lewis 19 Sewill Chris 9 Triss Linus 5 Nisul Curtis 15 Turkis Lois 4 Mois Kasey 3 Tasey Lorne 3 Norle Kirk 1 Tirp Lou 1 Oul Ted 1 Ked Luna 20 Nula Petra 9 Tekra Lyle 5 Nyme Titus 13 Kipus Mara 5 Rama Kathy 2 Thaky Marla 1 Marma Katia 1 Takia Megan 27 Negam Kate 49 Pate Miles 32 Mooles Tucker 27 Keeter Milo 15 Nilo Tate 18 Pake Mona 4 Lona Trista 2 Trispa Morris 3 Romis Terry 4 Reppi Moses 7 Somis Pierce 12 Kierce Myah 6 Lua Tracy 3 Satry Nathan 167 Thanen Carter 202 Tarker Noam 1 Loal Kipp 2 Keek Noel 9 Luel Kurt 6 Treek Noelle 11 Loenne Pippa 4 Teepa Norah 32 Morah Tessa 33 Seka Nya 6 Loa Tia 8 Eeka Owen 164 Owem Jack 198 Kaj Renee 9 Neray Greta 10 Tregga Ronin 20 Norin Victor 28 Tikver Rosanne 1 Noraz Yvette 1 Eetev Warren 8 Warrim Garrett 13 Garek Will 7 Wum Zach 6 Kaz

213

Appendix C

Stimuli from Chapter 4

Table C1

Word stimuli from LDT in Experiment 1.

Non-Onomatopoeic Iconic Onomatopoeic Iconic Arbitrary STOP PAT ROOST SILK FLICK CREED DREAD PEEP QUAIL WHIRL THUMP BIB SWIRL THUD MYTH SHRIMP MURMUR COUNSEL CRAWL SMASH REFUTE MASH HOOT SLACK TICK BUMP COPE SCRATCH BOOM INVENT GOO ZOOM CLOAK SWERVE BANG SHARE BREAK RIP WAKE TWIST BOUNCE SOLVE HEAVE CUCKOO BLAME SLIME RUSTLE FEE MUSH CHOP WAIL TWIRL TAP LAMP SLIP CRUNCH FLEE JOLT WHISPER GOOSE SUCK SIZZLE SOCK FLUFF GROAN SOAP REEK GROWL FIT SLUR TINKLE DUCK CLASH ZIP TURTLE BLAST CRACK NAIL PRICK SNARL VIEW FLASH CRICKET PEAR ITCH SCREECH SPOON GUSH SQUAWK BISCUIT DRUM CLANG MAMMAL THUNDER MOAN LIME TANG CRASH JEEP TRICKLE QUACK STEEPLE DRIZZLE RUMBLE ABIDE ZIGZAG MEOW AMBLE LIMP STOMP CHIPMUNK

214

WHIP BEEP PLATYPUS WADDLE CLUNK EEL FLICKER BLEEP HURTLE

Table C2

Pseudohomophone stimuli added for Experiment 2.

Based on Non-Onomatopoeic Iconic Based on Onomatopoeic Iconic Based on Arbitrary BIRST AWTER BUZZURR BLAWK BAUCE CHURP BRAUD BEARN CRACT CHYME CANSIR CRINKLEE CURLEE FITE HUMB EKO FLORE KLANCK GLAUCE KAFF KLAP HAWP KAWL KLIK HAYZIE KORT KOFF HOAL MANIDGE KOO MUSTEE NIFTEE KRINKLE SCURREE PULE KRISP SHAWK ROOLER NAWK SHUTE SAWKURR RORE SKREEM SHURT SHREEK SKUTE SIGAR SLIRP SMOAK SKURT SLUSHEE SNEEK TAWL SQUEEL STAIL TODE YON SWEAP ZOAN ZIPPUR