Summary and conclusions

What follows below is a chapter by chapter summary, a discussion of the relevance of the results, and some directions for future research.

Part I: Ideophones

Part I of the thesis introduced ideophones and surveyed earlier research. Chapter 2 captured the cross-linguistically robust features of ideophones by defining them as marked words that depict sensory imagery. After explaining the elements of this definition, much of the chapter was devoted to unpacking what can be thought of as its rich point: the claim that ideophones are depictions. As depictions, ideophones enable others to experience what it is like to perceive the scene depicted. I justified this claim by showing that ideophones are closely akin to physical demonstrations; that they are flagged as depictions in actual use; that speakers explain them in terms of “seeing as”; and that they show formal characteristics of depictive signs like relative iconicity and expressive morphology.

The depictive nature of ideophones is what underlies their sensory meanings and their uses as appeals to personal experience. It also provides a simple explanation for the markedness and aloofness of ideophones cross-linguistically: this is what signals the switch in mode of representation, helping to literally set apart the depictive ideophone from the surrounding descriptive material in the linear speech signal. Importantly, depiction is not to be equated with simple imitation or physical resemblance. Depictions make use of culture-bound, socially mediated representational conventions. It is this socially mediated nature that gives ideophone systems their language-specific signature.

Chapter 3 surveyed the history of linguistic work on ideophones, tracing the rise of ideophone research as a subtradition in African linguistics, the influence of writers such as Lévy-Bruhl in shaping the public imagination of ideophones as signs of primitive mentality, and the more recent cross-linguistic turn that led to wider recognition of the significance of ideophone systems. It emphasized the need to place ideophones in the context of the broader ecology of linguistic resources, and noted that the most blatant gaps in our knowledge concern their meaning and use in situated social interaction.

Part II: Siwu

Part II described ideophones in the context of the broader linguistic ecology of Siwu, the richly ideophonic Ghanaian language that was the subject of the field research described in the thesis. Chapters 4 and 5 provided the ethnographic and linguistic backdrop to later chapters, with chapter 5 providing a sketch of the grammar of Siwu and showing how ideophones fit in the broader systems of property-denoting expressions, the language of perception, and quotative devices. If this chapter showed that ideophones are well integrated in the broader linguistic system, the role of chapter 6 was to demonstrate that they nonetheless form a clearly distinct word class.

Findings of chapter 6 include the fact that ideophones are structurally marked by means of their word length, deviant phonotactic patterns, peculiar word forms and expressive morphology; that expressive morphology is not just random modification but exhibits an orderliness of its own and underlines the depictive character of the ideophonic sign; and that there is an inverse relation between the syntactic integration of ideophones and their susceptibility to expressive morphology and performative foregrounding. This latter finding illustrated the importance of corpus data in bringing to light possible lexicalisation or normalisation paths for ideophones. The chapter also addressed the common conflation of ideophones and interjections, arguing that despite some superficial similarities, these word classes are fundamentally different in semiotic as well as interactional terms.

Part III: Meaning

Part III examined aspects of the meaning of ideophones. Chapter 7 showed how different types of iconicity allow ideophones to move beyond the imitation of singular events towards perceptual analogies and generalisations of event structure. It also argued for the importance of the actual performance in establishing mappings between sound and sense, with a breaching experiment confirming that ideophones are produced as performances; and it explained the fact (often underplayed in the literature) that not all ideophones are transparently iconic by arguing that not iconicity, but the depictive mode of representation is fundamental to ideophony.

Chapter 8 probed six perceptual domains using dedicated stimuli, finding that ideophones are the sensory words par excellence. The results of this chapter do not only tell us about the language of perception in Siwu, they are also of wider relevance because they show that stimulus materials can be used to elicit ideophones and can play a key role as a tertium comparationis in developing a semantic typology of their meanings.

Chapter 9 used folk definitions to gain insight into the meanings of ideophones. It showed that such definitions specify important aspects of the background knowledge against which ideophones are understood, and that depictive gestures are crucial to illustrating the imagistic meanings of ideophones. Moreover, speakers often employ other ideophones as semantic anchoring points in the definitions, revealing sense relations of similarity and dissimilarity (but not, significantly, hyponymy or hyperonymy). This latter point suggests that all ideophones operate at a similar level of specificity, something that again meshes well with their status as depictions of specific sensory imagery.

Chapter 10 used a sorting task to tap into the conceptual structure of the domain of ideophones. It showed that speakers handle ideophones consistently and that for them, the domain is organised by richly diverse aspects of sensory perception, from SURFACE APPEARANCE to SPATIAL EXTENT and from MOUTH-FEEL to TEXTURE. This represents an important improvement over previous semantic classifications, which hitherto have been based on analyst’s intuitions more than on native speaker judgements. The chapter thus showed that a sorting task can provide a way to explore the semantic fields of ideophones without imposing preconceived categories.

Part IV: Use

The greatest blank in previous research has no doubt been the lack of studies of the ideophone in actual use, and especially in naturally occurring contexts. This has hampered progress on a number of fronts, and the chapters in Part IV addressed some of the questions that can be answered with this kind of data. Most important, in terms of setting the record straight, is the finding that ideophones are far from the erratic stylistic flourishes that they have been made out to be. In chapter 11 I showed that speakers wield ideophones as communicative precision tools in everyday speech, using them to share in sensory perceptions and to subtly sort out matters of epistemic authority. I was able to explain these uses with reference to work done in earlier parts of the thesis: they build on the depictive nature of ideophones (chapters 2 and 7) and their close connection to sensory imagery (chapters 8-10). The chapter focused on naturally occurring, informal talk-in-interaction because this is the basic stream of verbal behaviour that underlies all other ways of speaking. As we saw, earlier proposals about the use of ideophones run into problems as soon as we leave the familiar territory of narratives: they are either too restricted or too sweeping to account for many observed patterns of use. In contrast, the data-driven analysis of this chapter easily incorporated all earlier proposals while at the same time offering the explanatory power to accommodate the new findings.

Chapter 12 extended the investigation to two special ways of speaking: greetings and funeral dirges. In the context of received views about ideophones as quintessentially spontaneous, dramatic and idiosyncratic, the common use of ideophones in these genres might be thought of as unexpected. As we saw, however, it is not so unexpected after all: the baseline established in the previous chapter allowed us to see that it builds on some of the core interactional functions of ideophones in everyday speech. The use of ideophones in these disparate genres points to a common element that underlies them both: an emphasis on the sharing of experience that values communality and being together. The chapter also drew attention to the way in which the aesthetics of everyday language —in the form of ideophones— may feed into genres of verbal art.

Chapter 13 documented for the first time some clear cases of ideophone creation in a corpus of naturally occurring speech. An analysis of these cases showed that ideophone creation relies on general principles outlined in earlier chapters, for instance the use of performative foregrounding to signal a switch to the depictive mode of representation, and the use of the regular iconic mappings outlined in chapter 7 to suggest meaning. Ideophones are thus not created out of thin air, but build on the communicative competence —including ideophonic competence— of speakers of Siwu. One implication for the issue of language and creativity is that this form of linguistic creativity is based not on generating novel meanings using connotations and denotations of existing words, but instead on presenting new verbal material in such a way that the interlocutor treats it as a creative depiction.

Chapter 14 used corpus data to examine previous claims about the link between ideophones and gestures. It found that these claims are too strong: it is not the case that ideophones are almost always accompanied by gestures. Moreover, discourse type makes a difference: ideophones are more commonly accompanied with gestures in tellings than in other contexts. The chapter also found that there is a special affinity between ideophones and depictive gestures. Based on these findings it rephrased a conjecture by Kita to say that “if a communicative move features both a depictive gesture and an ideophone, the two will tend to be synchronised.” Finally, it proposed two reasons for the tight coupling of ideophones and depictive gestures: first, the fact that both ideophone and gesture are holistic depictions of complex states; and second, the fact that both form part of a single performative act.


A good deal of the work done in this thesis can be understood within a larger framework of putting the study of language in its proper context. This means not only looking at language in situated everyday interaction, but also taking stock of the ways in which the semiotic affordances of speech are exploited in the world’s languages. From this perspective, it is merely a historical accident that the study of language has focused for so long on isolated sentences and monologic texts, foregrounding a view of language as a vehicle for the disembodied and decontextualised transportation of ideas, and backgrounding issues of socially grounded meaning construction, the management of social relations, and the depictive affordances of verbal material. As we saw, ideophones thrive in everyday social interaction. To ignore them is to miss out on a rich cultural meaning system that forms an integral part of the communicative competence of speakers of ideophonic languages.

One conclusion to be drawn from the material presented in this thesis is that depiction is not a marginal phenomenon, but one of the central affordances of language. This should not surprise us. Whereas a language engineer might do well to take the arbitrariness of the sign as a design principle, natural languages are the result of aeons of cultural evolution in the hands and minds of language users. It would be rather more surprising if this semiotic affordance of verbal material had remained unused in human language. In language as in any cultural tool, “the semiotic and the material constantly cross-cut and convert into each other” (Sutton 2006:92). Repeating an observation made in other chapters, what is typologically interesting about ideophonic languages is that in them, the depictive potential of speech has taken on a life of its own, in the form of a sizable class of words which makes use of it. We can think of this as one of the many possible trajectories of the ever-evolving bio-cultural hybrid that is human language (Keller 1998; Croft 2000; Evans and Levinson 2009). This trajectory has often been overlooked or downplayed by linguists working on languages that lack such word classes. In this thesis I have shown that depictive words can be honed into communicative precision tools that enable speakers to evoke detailed sensory imagery and to manage matters of epistemic authority. So ideophone systems illuminate another corner of the design space of human language, and offer an important corrective to our ideas of what is possible in language.

Some big questions about ideophones have been why-questions: why are ideophones marked, why do they depict sensory imagery, why do languages have them at all? Such questions are notoriously difficult to answer, but in this thesis I have tried to show that reframing them can be rewarding. For instance, few explanations have been put forward for the fact that ideophones tend to be marked, performatively foregrounded and apart from the rest of the utterance (but see Kunene 2001; Nuckolls 1996). In chapter 2 I showed that we can turn the matter around and ask, given a stretch of verbal material, what is the best way to signal a switch in mode of representation? The answer is: by foregrounding some part of it so that it stands out as different and can be interpreted as a depiction. This foregrounding is precisely achieved by the marked forms, expressive morphology, and prosodic conspicuousness of ideophones. Another question is why cross-linguistically, the meanings of depictive words seem to be limited to sensory imagery. In chapter 7 I showed that if we turn the matter around and ask what can actually be depicted in speech, it becomes clear that there are important similarities between speech and sensory imagery which make it much easier to depict sensory imagery in speech than, say, objects or spatial relations. Similarly, some have wondered about the use of ideophone systems. After all, if they are merely “playthings” (Müller 1895) or “response cries” (Pinker and Jackendoff 2009) it is rather difficult to see the purpose of maintaining such large inventories. Again, reframing the issue helps: by shifting the question from “why do languages have ideophones” to “what do people do with ideophones?” we were able to see in chapter 11 that far from being mere embellishments or expressive outcries, ideophones are tools used for interactional work.

This reframing reflects a fundamental choice of perspective that underlies the work reported in this thesis. Evans-Pritchard once mused that ideophonic languages are “not so much a type of language as a revelation in language of a type of mentality” (1962:145). Along similar lines, Nuckolls (2004; 2010b) has proposed extralinguistic and cultural factors that may foster (as in the case of animistic cosmologies) or inhibit (as in the case of abstract and logical styles of scientific discourse) the development of ideophone systems. While such factors may be at play, in this thesis I have refrained from hypothesizing at all too abstract levels of causality. Instead I have built my case from the ground up as it were, starting from facts of form and function and from nitty-gritty details of talk-in-interaction. In this way I have been able to show that Siwu speakers use ideophones to share in sensory perceptions and to sort out matters of epistemic authority; and how these uses are grounded in, and fitted to, the nature of ideophones as marked words depictive of sensory imagery. We need a large cross-linguistic database of “micro”-level findings based on rich primary data like this before we will be able to talk more meaningfully about “macro” factors that may be at play in the development and use of ideophone systems.

Future directions

This thesis makes a case for the relevance of ideophone systems to the science of language. Its findings speak to several issues, from the typology of ideophone systems to sound-symbolism and from brain-imaging studies to the study of verbal art. Here I point to some future research directions.

Ideophones have often simply been characterised as iconic words. In chapter 7 I put forward a more nuanced view, arguing that it is not iconicity (resemblance) per se but rather the depictive mode of representation that is crucial. From this follow two points that could be fruitful loci for future research. First, if, as I have argued, performance is crucial to the interpretation of depictive material, this has implications for research into sound-symbolism. So far, experimental work on sound-symbolism has given little thought to the factor of how stimulus items are presented. One hypothesis following from the arguments presented here is that the manner of presentation actually matters a lot for how participants interpret the items presented. One prediction is that people should be more prone to form iconic mappings for stimulus items that are presented as performances (§7.5.1) than for items that are presented following standard protocols.

A second issue relates to the typology of ideophone systems. I pointed out that not all ideophones are transparently iconic, and I put forward the hypothesis that ideophones with less transparent form-meaning mappings (which I compared to “artists’ renditions” in §7.5.2) can only exist in virtue of depictive signs that show more concrete iconic grounding in reality as apprehended by our senses. The hypothesis can be couched in terms of an implicational hierarchy motivated as follows. If we are going to use speech depictively, the most natural first step would be to depict sound with sound (imagic iconicity). A natural next step would be to tap into the fact that auditory events often come packaged together with movement and perhaps intensity. This is where devices like Gestalt iconicity and relative iconicity come in. From there, a natural next step is to take advantage of the fact that many sensory events can be served by Gestalt and relative iconicity even if they do not involve sound, because they do share suprasensory attributes like aspecto-temporal unfolding and intensity. That state may finally function as a springboard to extend the idea of depiction even to inner feelings, sensations and cognitive states. The important point is that all these states naturally follow from each other, and the latter states are not likely to arise without the former being in place. This, I propose, is why we find ideophone systems that are limited to sound only, or to sound+movement only, or systems spanning the broader spectrum of sensory imagery to various extents, but why we do not find ideophone systems that are limited to inner feelings or cognitive states only. Although I am not aware of counterexamples, this proposal for an implicational hierarchy is in need of wider cross-linguistic scrutiny.

Another open issue concerns the differential elaboration of sensory modalities in ideophones. We have fragmentary reports —for instance, Japanese has been said to be richer in ideophones evoking cognitive states than African languages (Childs 2001:70)— but the issue awaits structured comparative typological investigation. The evolutionary sequence described above is just one limiting factor in what will no doubt turn out to be a complex interplay of many other factors, including the natural environment, subsistence style, material culture as well as language ideologies (Nuckolls 2004) and cultural and linguistic differences in the relative weighting or elaboration of sensory modalities (Geurts 2002; Majid and Levinson 2011). We don’t know much yet about the contribution and interaction of these different factors, but most importantly, we lack the rich primary data that would allow us to at least construe a picture of cross-linguistic variation in ideophone systems. Some of the methods used in this thesis, for instance dedicated stimulus materials, folk definitions and sorting tasks, can help out here.

The need to construe a more complete picture of cross-linguistic diversity naturally extends to the uses made of ideophones. Part IV of this thesis has identified a number of recurring patterns of use for ideophones in Siwu. Do we find similar uses in other ideophonic languages? No doubt there are uses of ideophones not represented in the corpus that informed this thesis. How do these relate to the nature of ideophones as marked words that depict sensory imagery? The only way to find out is to see how speakers of ideophonic languages across the world use ideophones in talk-in-interaction. A world of exploration awaits us.

Since ideophones employ a different mode of representation (depictive rather than descriptive, §2.5), there is reason to expect that this may involve processing differences that can be investigated using psycholinguistic methods. One recent study (Osaka and Osaka 2005) has suggested that Japanese ideophones for laughter, but not nonsense syllables, activate striatal reward areas in the brain. However, since the baseline is with nonce words and not with non-ideophonic words related to laughter, we cannot be sure that this can be attributed to the ideophonic status of the words. The effect may just be an instance of more generally known effects of embodied semantics (Pulvermüller 1999). At the same time, as shown in Pulvermüller’s work and in many follow-up studies since, there is clear evidence for the importance of sensory imagery in mental representation, and for processing differences between different word classes. Foroni & Semin (2009) showed that English verbs for laughing (smile, frown) and corresponding adjectives (happy, angry) induce motor resonance in facial muscles, the verbs more strongly than the adjectives. My hypothesis would be that ideophones for emotional states and facial configurations would rival verbs in the extent to which they cause motor resonance. Clearly, ideophones —with their tight links to sensory imagery and non-arbitrary mappings of sound and meaning— provide an ideal testing ground for questions about the importance of imagery in language and cognition. They have so far escaped comparison, probably simply because cognitive science has until recently been hardly aware of their existence. By supplying linguistic and behavioural evidence for the tight link of ideophones and perceptual knowledge, this thesis has done some of the groundwork needed to begin addressing these issues.

A final matter for future investigation lies in the study of verbal art and its relation to ordinary language. As I argued in chapter 12, many of the literary notions and analytic categories employed in the study of verbal artistry —notions like aesthetic involvement, enactment, foregrounding, multimodality— must in the final analysis have their root in the affordances of speech in everyday social interaction. The analysis of verbal art often proceeds in isolation from the everyday linguistic practices of a community, as if artful language belonged in a glass showcase. Yet to isolate verbal art from everyday talk is to ignore its crucial dependence on the aesthetics of everyday language. The contrastive analysis of mundane talk and a special genre like the funeral dirges discussed in chapters 11 and 12 demonstrated this by showing one way in which the aesthetics of everyday language —in the form of ideophones— feed into verbal art.

Ideophones have variously been reduced to playthings or upheld as highly poetic. Either perspective fails to see the role of mundane everyday social interaction in constructing our social world through linguistic resources. The research reported in this thesis gives new meaning to the words of Evans-Pritchard, who wrote half a century ago that ideophones are “poetry in ordinary language” (1962:145). Throughout this study, we have seen how the threads of the everyday and the aesthetic interweave. With both poetry and ordinary language in mind, the future of ideophone studies holds much promise.

Zie ook de samenvatting in het Nederlands.