Note to readers: A version of this argument has been written up and published as:
Dingemanse, Mark. 2015. “Ideophones and reduplication: Depiction, description, and the interpretation of repeated talk in discourse.” Studies in Language 39 (4): 946–970. doi:10.1075/sl.39.4.05din (PDF).
Depiction is a technical term used in psychology1, philosophy2, and art history3, but less so in linguistics4. One of my claims about ideophones is that they are best understood (typically, canonically, with the customary provisos about the fluid categories of natural language) as depictive words. Do we really need a term like that? Can’t we make do with good old ‘expressive’ or ‘vivid‘? No, I say. But what then is this elusive concept of depiction? How can we tell that something is intended as a depiction? What is depiction, such that we can distinguish it from other modes of representation? In answering these questions, it is useful to make a distinction between the signals that frame something as a depiction on the one hand, and the aspects of depiction as a mode of representation on the other hand. But let’s start with something more basic: words and images.
Words versus images
There is a domain where distinct modes of representation are very important and very clear: that of visual representations, in particular written words versus images. Take the following slide (from one of my presentations), with on the left a sentence involving le soleil “the sun” and on the right an image of the sun. (Bonus points for those who see whose handwriting that is.5) The claim here is a simple one: that the left and the right side involve two distinct modes of representation.
I’m guessing most readers will agree with at least this basic distinction. What we call this distinction doesn’t matter a lot, but I call it, following the literature, “description” versus “depiction”. More important than the labels are the ways in which the phenomena differ. On the slide above, a first opposition is between propositional vs. imagistic. This is in a way restating the same point (see Kosslyn 1980 for more details), so I’ll focus on the remaining three. There are three key differences that help us to deductively distinguish depictions from descriptions: descriptions differ from depictions (1) in terms of symbol system they use (using discrete symbols vs. using gradient markings), (2) in terms of form-meaning mappings (basically arbitrary vs. basically iconic), (3) in terms of how we interpret them (“decode” to interpret vs. “imagine” to interpret).
Words as images
Without even having started worrying about ideophones, this is a distinction that we need to make. It’s a distinction that is motivated on independent grounds. Words are different from images in their mode of representation, and we find a number of clear differences between the descriptive and depictive modes of representation.
The next step of my argument consists simply in noting that precisely those three key ways in which descriptions differ from depictions (in the visual mode) are also ways in which plain words differ from ideophones (in spoken language). Ideophones tend to use more gradient and discrete symbol systems; their form-meaning mappings tend to be more iconic than arbitrary; and to interpret them, we “imagine” more than we “decode” (Dingemanse 2011, 2012). Of course these differences are not absolute; in my thesis I have pointed out, for instance, that convention (and hence decoding) also plays an important role in depiction. After all ideophones are conventionalised words and not creative formations — most of the time.
When I say that ideophones are depictions, this is a statement about their mode of representation. It is more about a way some thing can be than about the thing itself. Take the sun. That’s a pretty good ‘thing’. Now take the two representations of the sun shown above. Even those representations are tangible ‘things’, in the sense that we can talk about them, point at them, isolate them (as I do on the slide above). But the way in which those two representations differ, that is their mode of representation. It is about the way something can represent something else. That is the sense in which I call an image of the sun and an ideophone in spoken language a depiction.
Signals and aspects of depictions
Having established this sense of depiction, we can next make a distinction between signals and aspects of depictions. A signal serves to alert the interpreter to a difference. It’s like the frame of a painting. That frame is not itself the depiction, but it frames it, marks it as distinct from the surrounding material. Distinct how? In mode of representation. Good examples of signals of depiction in speech are things like quotative markers or performative foregrounding. Both can be seen in an example like “The car went VROOOM” — the ‘go’ verb functions as the quotative marker, and the capitals of vroom help highlight its performative nature (a convention familiar from comic strips). Note that there is nothing depictive about the quotative verb itself. It just serves to introduce someting as as depiction.
Aspects of depictions is what I want to call the more direct manifestations of the depictive mode itself. For instance the smudges of paint on a painting, dark in some places and light in some others, thick in some places and thinner in others. This gradient use of paint is what makes the painting a depiction rather than a description. Good examples of aspects of depiction in speech are the iconic form-meaning mappings we find in ideophones: for instance the fact that word structure may resemble event structure in ideophones like dzâ! ‘sudden appearance’ or fwɛ̃fwɛ̃ ‘flexible’ (the technical term for this is Gestalt iconicity, Dingemanse 2011). These forms of iconicity, found commonly in ideophones but rarely in ordinary words, show that we find ourselves in another mode of representation, one in which verbal material is being used in gradient ways.
Of course, semiotics being what it is (or more to the point, we humans being what we are), signals and aspects may sometimes be conflated. The use of paint in a painting is not only an aspect of the depiction that helps you imagine the scene; it also serves as a signal that alerts you to the fact that you are now looking at a depiction and not a description. In language, performative foregrounding is not only a signal that we are now listening to a depiction: it can also serves as an aspect of the depiction, as in Janis Nuckolls’ descriptions of Quechua ideophones where intonation is used to signal height and other things (Nuckolls 1996). And in an ideophone that is reduplicated eleven times in my Siwu corpus, the expressive morphology serves not just as a signal that we’ve entered the depictive mode, it is also an aspect of the depiction itself in that it may iconically signal ongoing repetition.
The burden of alternative accounts
In closing, it is useful to think about what this account is giving us — and conversely, how we could argue against it. What it gives us is a unified explanation of a set of features that have struck linguists as noteworthy from early on. Ideophones have always been described in special terms such as “vivid”, “expressive”, “playful”, “peculiar”, but none of these terms have been properly defined and so we’ve been stuck, for a long time, with subjective, piecemeal characterisations of ideophones. Here I have argued that the depictive mode of representation is a key feature (perhaps the key feature) distinguishing ideophones from ordinary words.
If one really wanted to argue against a depictive account of ideophones, one would have to come up with good reasons for the conglomerate of signals and aspects that I argue point us to the depictive nature of ideophones. One would have to explain why they often come with quotative marking; why their forms are more gradient and iconic than ordinary words; why they are in many languages susceptible to expressive morphology; and why they are often performatively foregrounded. That seems a tall order. This is precisely the thought process I went through before ending up at the solution presented at length in my thesis and summarised here. The solution lies in recognising that ideophones differ from ordinary words in terms of mode of representation, in a way that is analogous to the difference between written words and images.
Note. I thank Kimi Akita and Herb Clark for their thoughtful questions about these matters, which over the years have helped me to clarify my account of ideophones as depictions.
17 responses to “Description and depiction”
Don’t forget the diagrammatical iconicity aspect- while not all ideophonically rich language systems appear to make strong use of this (esp. in the mapping of meaning to form), many languages do seem to. That some don’t had been a mystery to me at first, then I realized that may have to do more with the form class orientation of a language. When we talk about morphosyntactic type, we usually limit ourselves to normal lexicon and morphology. But ideophones can be the third leg of the hierarchy, as (ideally and prototypically) prelexical as grams are postlexical once fully grammaticalized. Not all ideophone systems make equal use of excess prosody or gesture- that is there is a cline to the degree of imagistic iconicity. And I’m betting that that more phonosemantically domesticated ideophone systems are a later development away from the more purely imagistic state. It would be interesting if there are also morphosyntactic concomitants.
Oh Jess, for once I abstract away from all the finegrained typological detail to get a clear focus on the distinction between prototypical depiction and description, and then you come pouncing on these points! ;)
To be clear:
(1) Gestalt iconicity (along with relative iconicity) to me is precisely one of those types of diagrammatical iconicity that go beyond mere imagic iconicity (Dingemanse 2011).
(2) Not all ideophones are equally depictive, a point underlined by corpus and typological data in my paper on Expressiveness and system integration (preprint).
(3) There is a cline, but not just in terms of iconicity (a position held more strongly by Kimi Akita I think). I call it an implicational hierarchy; more details in Dingemanse 2012 (pdf).
So yes, I agree with you. ;)
Thanks for the post, Mark!
1) I’m interested in what kind of boundary you posit for the ideophone category. Judging from your definition of ideophones (i.e., “marked words that depict sensory imagery” [Dingemanse 2011: 3]), which doesn’t make use of a proviso or an adverb like “typically,” you seem to think of a clear (i.e., non-fuzzy) boundary like the ones illustrated by the mathematically definable category of ODD NUMBER and perhaps by the biologically definable category of BIRD (Rosch 1973; Fillmore 1982; Armstrong et al. 1983). But I’m not sure if depiction can serve as such a necessary and sufficient condition of ideophones (though I guess it doesn’t need to in your theory).
2) I was wondering if the degree of ideophonicity can be metaphorically captured as that of realisticity. Both realistic and abstract paintings are depictions, but they differ in how realistic they are (I don’t know how different this cline is from my iconicity hierarchy). Just another elusive thought.
Ah your questions prod me to write another post — now on categories, boundaries, and comparative concepts. But in brief, no, I absolutely don’t think of ideophones as a rigid category, and of depiction as a necessary and sufficient condition. Once again, the fact that I focus here on getting the depiction/description distinction clear necessitates cutting some corners and leaving out some of the typological detail. In fact I think more of ideophones as a category with prototype structure à la Rosch; or in more recent typological terms, as a description that defines a canonical prototype.
From my thesis, p. 34: rather than stipulating necessary and sufficient conditions for ideophonehood, the definition characterises a general prototype and in doing so provides a reference point for discussions of language-particular implementations.
Ideophones are not very different from other categories in this respect. Language is a fluid, evolving phenomenon. Its categories are not set in stone and will show the hallmarks of culturally evolved entities: fuzzy boundaries, prototypicality, networks of historical relations. Take any linguistic category and you’ll see that you’ll need a proviso or hedging adverb like ‘typically’. It would be silly to build that into the definitions; rather, we should pragmatically use definitions for what they are: characterising the core characteristics of a phenomenon without getting hung up over the fact that there will be a fuzzy periphery. Indeed the basic idea of canonical typology is that if you lack a clear definition of this core concept, you’ll have trouble understanding exactly what is fuzzy about the periphery.
Does that clarify?
As you know I flank lexicon by postlexical grams on the one hand and prelexical ideophones (as ‘antigrams’). Historical relations show a general trend towards slow evolution prelexical>lexical>postlexical (plus an apparently much faster and more conscious reverse trend which can turn lexemes, modified or otherwise, into ideophones (often with truncation), or grams into lexemes (often with additions). But the categories themselves have very fuzzy boundaries as we are discussing. Any particular entity can add or delete relevant features a bit at a time (or even split ending up with different feature sets (maybe merge losing them?). Is there some metric for dealing with ideophones from this perspective the way some manage to deal with generalized lexical forms canalizing into grams?
I agree that ideophones constitute a fuzzy prototype category. The point was that not all prototype categories have a fuzzy boundary. For example, even the ODD NUMBER category, which is mathematically defined and therefore clearly delineated, shows “prototype effects”: 3 is a better example of an odd number than 3123 (Armstrong, Gleitman, & Gleitman 1983). I once attempted to draw a clear category boundary around Japanese mimetics, which appear to be primarily defined by a limited set of morphophonological templates (reduplicative, suffixal, etc.). But the existence of some words with a typical mimetic shape but with unclear mimetic status (e.g., zut-to ‘all the time’, sokkuri ‘exactly like’) led me (back) to the “fuzzy” view.
Is the category being fuzzy, or are we just being lazy intellectually? We can be featurally nitpicky when we want or need to be. So we have phonologies ranging from something like 7 up to 150 or so (but depending on definitions here also). Prototype hierarchies seem to get created when such fine detail becomes overbearing, and combinatorics takes over from unique forms to save effort. If we look at ideophone hierarchies, it just seems to me, based on inspection of inventories (some just sampled by lexicographers), that the imagics/imitatives are the last to get lexicalized. Also they tend to be most of those present when an inventory really is very small. Sound trumps visual sense, but both trump all others, as well as abstracts. The flip side then is that the others tend to lexicalize most easily. So perhaps something about distance? The more intimate, the more lexical? There does seem to be something 3rd personish about ideophones.
That’s a false dichotomy if there ever was one. I think insisting on crisp category boundaries and looking away from the areas where one phenomenon shades into another actually comes closer to intellectual laziness. Saying that a category is not defined in the classical sense, with necessary and sufficient conditions, does not remove the need to be precise about the nature of the phenomenon we’re defining. I think typologists have found excellent ways to deal with the culturally evolved categories of natural languages — away from classical Aristotelian categories and Platonic ideas, it is true, but taking more seriously the way linguistic categories evolve and interact synchronically and diachronically.
As to your other points, Jess, when are you starting a blog where you put together all your data, observations and hunches? It would be great to have all this in one coherent place instead of distributed over so many threads on the interwebs.
@Kimi, it seems we’re in full agreement then. I probably misunderstood your first point because, prototype effects notwithstanding, it strikes me as highly implausible that any lexical category in natural language would be as strictly definable as the set of odd numbers.
I still wonder whether you really think that our definitions should always include a proviso like ‘typically’. Or was that just a clever provocation to get me on my hobbyhorse?
Well, I don’t know about one metric, but both Kimi and I have been concerned with the various graditions of morphosyntactic integration of ideophones. In my ‘expressiveness and system integration’ paper you can find some ideas about how to typologise ideophone constructions within and across languages along these lines. I think that comes close to your questions about lexicalisation/grammaticalisation.
As for your term ‘antilexical’, while I get the gist of it I think it might not be the most productive way to think about ideophones. If our sense of lexical includes sth like the notion of ‘being conventionalised’ (for me it does) then ideophones are clearly lexical. Corpus evidence suggests that at most 5% of ideophones tokens in natural conversation is not conventionalised. Those forms you could call antilexical or creative — but most are really conventionalised words with stable, definable meanings. The crucial difference between ideophones and ordinary words, to me, is not lexical vs. antilexical but, as you can guess from this post, their mode of representation.
I don’t say antilexical- but pre/postlexical in terms of historical development or hierarchically. Prelexical implies not having full lexical insertion privileges or whatever they call them these days, and postlexical implies having greater than normal id. From the Peircean perspective I like to think that grams have much in common with indexes (or indicies?). Lexemes would be prototypically symbols. And ideos would be icons diagrammatically, on the one hand, and imagically, on the other (taking two slots of four). Image would concentrate on performance representation, but with diagram we have more encoding. All cyclic, following morphosyntactic change/type. Just a hypothesis about statistics- probably plenty of exceptions but then we have that for typology generally already. Maybe I’ll publish when I can get time away from my current primary project on atomic Pascal math (which by the way goes very well, was just able to convince folks who deal with atomic clusters).
Ah yes, sorry I misread your term. Postlexical makes me think of Diffloth’s position in his (1980) Expressive Phonology and Prosaic Phonology in Mon-Khmer:
Well, I think I’m just curious :) I usually put “(proto)typically” on my semantic characterization of mimetics/ideophones. But it’s simply because I don’t want to spoil my discussion time by “excusing” the “exceptions.” I guess you’ve also experienced many Q&As with questions like “Is this an ideophone or not?”
FYI: Another boundary-related issue frequently discussed in Japanese linguistics is the distinction between onomatopoeic vs. non-onomatopoeic mimetics/ideophones (i.e., phonomimes vs. phenomimes & psychomimes). (In my opinion, the distinction itself is not central to mimetic semantics.) Noguchi (1995), a dictionary of Chinese onomatopoeia written in Japanese, says onomatopoeia is quite clearly distinguished from non-onomatopoeic ideophones in the language (p. ix). Because some researchers don’t accept the existence of non-derived non-onomatopoeic ideophones in Chinese, Noguchi’s (non-native) intuition of onomatopoeicity could be that of ideophonicity. Future research in ideophone typology might reveal different degrees of fuzziness for ideophones across languages, though such an investigation may have to follow the discovery of a reliable metric for ideophonicity or “(post-)lexicality.”
Haha, yeah, we could easily draw up a list of Frequently Asked Questions about ideophones, and this would be among them. Depends a lot on the audience though. Africanists are much more familier with whole word classes of ideophones and so their questions are different than those of ‘general’ linguists.
I really like this discussion.
I like to compare ideophones to Scott McCloud’s picture plane. He basically has a triangle with realistic art at one corner, abstract art at another, and comic (diagrammatic) art at the final corner, with that line extending words. I think phonomimes are the equivalent of abstract art; phenomimes of realistic art, and psychomimes of diagrammatic art. All of which I consider to be ideophones or mimetics.
So, it is not just a bi-polar continuum but rather a tri-polar continuum.
Ah but that’s a funny triangle though, with the extra leg (essentially an extra dimension) accomodating text. Whatever the degree of iconic abstraction on the lower leg of the triangle, there is still something of a resemblance on both sides (it’s aspects of the depictions that are different). The addition of text to that continuum, in the next step, obfuscates the description/depiction distinction, but is not so gradual (not sure whether I’m prepared to say it’s discrete, but processing-wise it is fundamentally different).
As he writes himself, “Cartoons and written language stare over a fence at each other on the right end.”
I understand the esthetic need to have it all in one triangle but I think that comes at the price of befuddling the description/depiction distinction. Which is totally fine of course — every framework and diagram is created for a purpose. Insofar as his purpose is to represent all forms of visual communication in one plane, this is basically the best he could have done. I quite like it as a way of depicting the possibility space, but it doesn’t mean that there is a neat tri-polar continuum.
I do like this point though:
That is very much in line with some of my proposals about how different types of ideophones depict in different ways (and are iconic to different degrees) — see e.g. my Ezra Pound among the Mawu paper, the implicational hierarchy of ideophone systems set out here, and my SemiotiX article on coerced iconicity.