Update: SemiotiX issue XN-8 features a revised and expanded version of this essay.

The linguistic blogosphere featured some posts recently on the topic of phonosymbolism, phonosemantics, and Chinese characters. It started with a post by Victor Mair over at Language Log, outlining several approaches to “etymologizing” Chinese characters. A follow-up by David Branner highlighted some of the problems with simplistic notions of phonosymbolism. Here I add some texture to the conversation by discussing the views of Ezra Pound, making a comparison to form-meaning mappings in ideophones, and introducing the notion of coerced iconicity.

The posts by Mair and Branner address a popular but quite mistaken notion: the idea that Chinese characters are like little pictures whose meaning can be “read off” from the strokes. The academic best known for debunking this popular misconception was John DeFrancis in his (1984) The Chinese Language: Fact and Fantasy. He showed that the bulk of Chinese characters are phono-semantic compounds in which one element indicates (at most) a general category of meaning and the other suggests the pronunciation.

The pictorial view

The roots of the “pictorial” view of Chinese characters in the Western world no doubt go far back. One of the driving forces behind it in the first half of the 20th century was the poet Ezra Pound. Pound is a fascinating figure, famed for his influence as a Modernist, Imagist poet and literary critic (and controversial for some of his political views). I have recently described Pound’s ideas about Chinese ideograms:

Over the years, Pound developed a fascination with the poetic affordances of logographic writing systems, especially Chinese. This fascination originated with his discovery of a theory of the Chinese character by Ernest Fenollosa [published in an edition by Pound in 1920], who argued that Chinese writing reflects etymology (‘true sense’) in a way that phonetic writing does not. In Pound’s idealist view of etymology (Li 1986), this rendered the Chinese character vastly superior to Western phonetic script in terms of picture-making. Soon enough however, scholarly studies of logographic writing systems showed that Chinese characters are semantic-phonetic compounds rather than transparent pictures, and Pound’s idyllic conception of Chinese characters as evocative ideograms was severely and justly criticized (Kennedy 1958; cf. DeFrancis 1984).

(Dingemanse 2011:44-45)

In my paper (titled Ezra Pound among the Mawu and published in Semblance and Signification), Pound’s ideas serve as a cautionary tale. I argue that there is a parallel between Pound’s overeager “iconicization” of Chinese characters and the tendency of many linguists to ascribe iconicity to ideophones. One important point of the paper is to note that there are limits to the iconic representational powers of speech, and there is reason to be careful in ascribing iconicity to ideophones (p. 45-6).

Ideophones are not the unproblematically imitative words that many people have made them out to be. There is something about them that makes us want to believe this, no doubt — just like there is something about Chinese characters that makes us want to believe the pictorial story. In my analysis of ideophones, this something is not iconicity, but first and foremost their depictive nature — the fact that they are presented as, or to use a more apt metaphor, framed as depictions.

Three types of form-meaning mappings

It may be useful to describe the development of my own thinking about these matters. Back in 2007 my reading of the ideophone literature suggested that ideophones are simply sound-symbolic words. Over time, with my inventory of Siwu ideophones steadily growing and my grasp of the semiotics of depiction in speech slowly evolving, I came to question simplistic notions of sound symbolism and iconicity in ideophones.

It became clear to me, for instance, that in a language with thousands of ideophones, it would be very difficult for all ideophones to be iconic to the same degree or in the same way. So there had to be different types of iconicity — different ways in which ideophones could evoke sensory imagery. My paper addressed this matter empirically by surveying the Siwu ideophone inventory. The result of this survey was a description of three basic, non-exclusive types of form-meaning mappings in ideophones.1

Coerced iconicity

While working on this I also realized that even if we allow for different types of iconic mappings, certain ideophones do not actually seem to be that transparently iconic. How does one iconically map colours, internal sensations, or cognitive states? Is iconicity really the point of ideophones like Siwu fùrùfùrù ‘seeing things in a blur’ or Japanese iya iya ‘with a heavy heart’? It seems unlikely. Have ideophone enthusiasts (native speakers as well as linguists) simply been over-eager in iconicizing ideophones? Doing an Ezra Pound in the domain of sound? If so, it is important to figure why the form of ideophones is so often identified with their meaning. I argue that it is their depictive nature:

Depiction, rather than iconicity, is what invites people to treat the ideophone as a performance of sensory imagery. An analogy may help to explain this point. Consider the category of objects called paintings. Paintings vary quite widely in the degree to which they are iconic (i.e. show a perceived resemblance to what they depict). And yet there is a distinct interpretive frame we bring to all of them: we tend to view them as depictions rather than read them as texts (Gombrich 2002[1960]; Walton 1973). In a similar way, we may think of ideophones as setting up a depictive interpretive frame, inviting the listener onto the scene and invoking images of being there.
(…)
If we want to invoke iconicity here at all, we should call it COERCED ICONICITY. The depictive nature of the ideophone coerces us into treating the word as an adequate rendition of the depicted event.

(Dingemanse 2011:51)

Coerced iconicity may be a useful concept in discussions of supposed iconicity because it describes a mechanism familiar to us all and realistically locates it in the eye of the beholder. In Peircean terms, it locates iconicity in the interpretants of eager observers rather than solely in properties of the sign-object relationship.2 Why was it difficult for Pound to resist associating meaning with the shape of Chinese characters? Why does the pictorial view of Chinese characters, thoroughly debunked as it is, keep coming back? One reason may be that there is some amount of truly pictorial characters that feed the imagination and that makes all Chinese characters look like pictures, especially to the untrained eye. This coerces people into treating all characters as pictorial renditions. Why do speakers treat all ideophones as perfectly adequate depictions of sensory imagery? Perhaps all that is needed is a critical mass of transparently iconic ideophones (using the three principles I described), and for the remainder, the framing devices of performative prosody and expressive morphology may be enough to coerce people into treating them as good depictions.3

Explanatory leakage

Sapir famously said that all grammars leak. Much the same holds for any grand theory of how linguistic signs —spoken as well written words— are motivated. (This is the source of my unease with the “big picture” theory of Chinese phonosymbolism by Howell that Mair outlines in his post.) All linguistic systems are the gloriously messy products of a long term interaction of human communicative needs, intersubjective language use, modality-specific features, and the mindless opportunism of evolution (among other factors). In the case of the form and meaning of ideophones, there are many forces tugging at them and shaping them. Although many people like to think of ideophones as prototypically “iconic” words, on reflection, it is clear that the story leaks. Yes, there are clearly iconic structures in ideophones that help guide the imagination, perhaps somewhat like the lines and shading in a naturalistic painting. But some ideophones (many in some languages?) may be more like abstract paintings: depictions that are invested with meaning by eager observers, not necessarily on the basis of information contained within their form.

Taking explanatory leakage serious is more realistic than sticking to a neat account and ignoring the periphery (Joseph 1997). Seeking regularity all the way tends to lead to oversimplification. In some possible world, all Chinese characters are neat pictograms and all ideophones are nice imitative words. That world is not ours however; and isn’t it is far more interesting to investigate the manifold ways in which humans can do cross-modal mappings of form to meaning, and to describe the different processes by which they discern motivation in what to the analyst may look like arbitrary gibberish? Gibberish. Hmm, let me frame that word for you so that you can experience some coerced iconicity on the way out. Gibberish.

References

1. DeFrancis, John. 1984. The Chinese Language: Fact and Fantasy. Honolulu: University of Hawaii Press.
2. Dingemanse, Mark. 2011. ‘Ezra Pound among the Mawu: Ideophones and Iconicity in Siwu’. In Semblance and Signification, edited by Pascal Michelucci, Olga Fischer, and Christina Ljungberg, 39-54. Iconicity in Language and Literature 10. Amsterdam: John Benjamins. (download here)
3. Dingemanse, Mark. (in press) “Advances in the cross-linguistic study of ideophones.” Language and Linguistics Compass.
4. Fenollosa, Ernest. 1920. The Chinese Written Character as a Medium for Poetry. Edited by Ezra Pound. London: Stanley Nott.
5. Joseph, Brian D. 1997. “On the Linguistics of Marginality: The Centrality of the Periphery.” Chicago Linguistic Society 33: 197–213.
6. Kennedy, G. 1958. ‘Fenollosa, Pound, and the Chinese Character’. Yale Literary Magazine 126, no. 5: 24–36.
7. Li, Victor P. H. 1986. ‘Philology and Power: Ezra Pound and the Regulation of Language’. boundary 2 15, no. 1/2: 187-210.
8. Pound, Ezra. 1947. The Unwobbling Pivot and the Great Digest. New York: New Directions.

1. The simplest of these, imagic iconity, maps sound onto sound. The other two are types of diagrammatic iconicity: in Gestalt iconicity, word structure maps onto event structure (accounting for the the fact that reduplicated forms often have iterative-distributive meanings), and in relative iconicity, relations between words (e.g. a difference in vowel quality) maps onto relations between meanings (e.g. a difference in magnitude). Further examples and discussion can be found in the paper and in Chapter 7 of my thesis. [↩]
2. Is not all iconicity then of the coerced type? No, I think that would be a step too far. As long as we find regular form-meaning mappings of the types described above, I think we may need some notion of lexical iconicity to fully understand them. It is just that we should not be greedy: there is no reason to expect all ideophones to be neatly iconic, just like we don’t expect all paintings to be naturalistic. [↩]
3. This raises all sorts of questions, by the way. What would the critical mass have to be like? Does it involve something like a minimum inventory size or an optimal relative weighting of the different types of iconicity? What is the relation between the different types of iconicity (and non-iconicity) and the change and development of ideophone systems? I have begun to address some of these questions in recent work, e.g. Dingemanse (in press) but much remains to be done. [↩]