What's a Word Worth is a new column by justira about the mechanics of writing. In this column, I examine the actual writing of every single book I read, focusing on how it conveys meaning and whether the writing works for me as an editor, reader, and fellow writer. My analysis will be based on the Peircian semiotic framework, explained in the first few posts of the column.
We're here! I'm finally going to talk about how metaphors work! Let's do this!
Welcome to the fourth post in my introductory series for this column. The first four posts will introduce readers to Peircian semiotics, which is the framework I use to analyze writing. If you're new to this column, please check the first, second, and third posts in this series. If you're following along live and are returning after a week away, you might benefit from taking a look at the review section of the previous post.
This is the last trichotomy, I promise! It's also the one with all the good stuff in it.
So!
There are actually three types of icons. Although iconicity has the potential to be of much interest to anthropologists, linguists have largely ignored the potential of icons, focusing largely on the symbolic aspects of language (see, for example, Saussure's dismissal of onomatopoeia and focus on symbols, no I will not let that go). While it is true that most of the lexical items — words — in language are symbolic in nature, a lot of the most interesting creative and social work of language is done through icons and indices(2). Icons have a powerful potential for social action because they are signs of possibility and similarity, and as was emphasized in previous posts in this series, what counts as similarity is culturally determined.
Of course, that is not all we can say about similarity, and in order to understand iconicity it is necessary to ask the question: what, precisely, do we mean when we say two things are similar? What kinds of similarity are there?
This is precisely the question that hypoicons — the three subtypes of icons: image, diagram, and metaphor — are equipped to answer.
The first question we must ask is: what does similarity tell us about the object? Recall that quite early on (in the second post) I said that iconicity is the only mode that can tell us anything about their objects simply by examination of the representamen. Recall also the importance of inference in the Peircian semiotic and the logical order of determination where an object determines a representamen (that's the post right before this one). Now we're ready to put it all together.
I can now substantiate my earlier claim about the information content of icons. Because an icon has inherent qualities that render it fit to serve as a representamen for some objects, we can, by examining a representamen that serves as an icon for a particular object, infer things about the object based on those very qualities that make the representamen a fitting sign for that object. It is in this sense that iconicity is the only modality that can give information, and we can see more clearly now why all signs have an embedded icon, even if it is "of a peculiar sort" (recall Peirce's definitions of the index and symbol quoted in the previous post). This informational relationship becomes especially evident when we look at the hypoicons in detail.
An image is a pure qualisign: it shares simple qualities with its object. Being a pure qualisign, an image has no concrete existence, and its presence is only perceived in existent objects like the redness of a rose or the pitch of a note. A linguistic example of an image is onomatopoeia, such as the sound of "buzz" or "purr" having a similar quality to the sound a bee or cat makes.
Linguistic representations of animal sounds, incidentally, serve as wonderful examples of the cultural relativity of similarity, as different languages come up with vastly different ways to represent essentially the same sounds, and claim that their particular representation is similar to the animal sound itself(3). While we might think "ribbit" is a fair approximation of the sound a frog makes, Japanese speakers think that "gero gero" is just as fair an approximation of the same animal noise and would probably reject our "ribbit" as not sounding much like a frog at all. But most onomatopoeic sounds across languages have some similarity to one another, all being images of the same sounds — such as the trill or flap in the sound we say a frog makes.
So, while an object may have many qualities, an image needs only one such quality to function as a sign for that object. In this sense, an image is monadic, a First Firstness (as icons are themselves monadic), independent of both interpretant and object. This situation has been illustrated in Fig. 4.1:
Figure 4.1: The first hypoicon, the image. The representamen is shown to have some qualities q1, q2, q3, while the object and interpretant are left unspecified, and no connections are drawn.
Note that the representamen has certain qualities q1, q2, q3 independent of the object and interpretant, which are left unspecified. The rabbit-duck illusion from the previous post is an image, but it has a little more going on besides that, which I'll cover soon.
A more involved linguistic example of an image can be found in poetry. Consider these lines from "The Raven" by Edgar Allen Poe:
And the silken, sad, uncertain rustling of each purple curtain Thrilled me—filled me with fantastic terrors never felt before;
The succession of sounds in the first line comes to suggest, through its phonetic qualities, the rustling of the silk curtains. The words themselves are symbols, but the qualities they share — whether or not we take them to refer to the actual sound silk curtains make — are there in and of themselves.
The representative character of the diagram, on the other hand, is tied to its dependence on the object. A diagram relies on analogous relationships between parts in both object and representamen; its capacity as a sign relies on there being at least two elements associated in a representamen that are likewise related in the object. Note that, just as with images (or any type of icon), the diagram's representational fitness is grounded by indexical relations that already exist in the representamen — in each case, the necessary constituents (qualities for the icon, relations for the diagram) must be present in the representamen. Diagrams, then, are dyadic. Such a sign is a map, where the distances and general scale of the map are proportional to their equivalents in the terrain the map represents. Note that most maps do not share simple qualities like colour with their objects (the red and grey roofs of houses that dot the landscape, the dark grey of asphalt, do not appear on most maps). An architectural blueprint is also a diagram; so are bar graphs and pie charts. Fig. 4.2 below is an example of the former:
In charts like this, the relationship between sizes of the bars or portions of bars is taken to be the same as the relationships between the relative sizes of the things they stand for. In this case, more awards were given to books by cis men, about cis men, than were given to books by cis women about anybody. Few awards were given to books by cis women about trans or nonbinary characters, and no awards were given to books by cis men about trans or nonbinary characters. This is shown by the relative sizes of the bars and portions of bars to each other. Elements of this chart are, of course, symbolic, such as all the legends that help us understand what the chart refers to. But the chart as a whole signifies the relative number of awards through a diagrammatic relationship between bar size and award number.
Going back to the rabbit-duck illusion, this also is in part a diagram because the relative sizes/positions of the eyes, head, and ears/beak match that of the object(s):
The rabbit-duck illusion again, which partakes of qualities of both an image and a diagram (source)
So those are visual example of diagrams. Feels pretty familiar, right? But there are linguistic examples, too. Perhaps because we are used to thinking of language in symbolic terms, iconicity is usually harder for us to perceive when it manifests in language as opposed to pictures.  But linguistic examples, while perhaps serving as poor explanations, are particularly illuminating once the basic idea is grasped. So, a linguistic diagram:
This surgeon is a golfer.
The diagrammaticity of this relationship can be illustrated in Fig. 4.3, with s="this surgeon" and g=the class of golfers:
Figure 4.3: The second hypoicon, the diagram. A relationship s—g is shown in the representamen and in the object, with a parallel connection drawn between the two.
Our subject is the surgeon, and we claim that there is a relationship between the surgeon and golfers: namely, that the surgeon is a golfer. The surgeon may also be a woman, a blues fan, or a dog owner, but all we are focusing on is that the surgeon is a member of the class of golfers. The membership relation between the individual surgeon and the class of golfers is what is common to both object and representamen. Notice that while the parts of the diagram itself require symbolic interpretation (so that we may know what is meant by "surgeon" and "golfer", for instance), the diagrammatic relationship itself requires no third, mediating entity: as it is in the object, so it is in the representamen.
Things are way, way different in the case of the metaphor (finally!), the last and most complex of the hypoicons(4). Peirce provides the following cryptic definition of a metaphor: the properties of the metaphor are such as to "represent the representative character of a representamen by representing a parallelism in something else"(5). This at first seems quite opaque (because Peirce is a blowhard), but I promise I will explain. Unlike the previous two categories, we will begin right away with a linguistic example:
This surgeon is a butcher.
The nature of this metaphorical form is illustrated in Fig. 4.4, with b=the class of butchers, m=meat and p=patients:
Figure 4.4: The third hypoicon, the metaphor. The object and interpretant show relationships b—m and s—p, while the representamen shows: s—b. This relationship is not directly present in either object or interpretant. The lines of association cross between representamen and interpretant, indicating that to make sense of the metaphor, we must untangle the parallelism using sociocultural knowledge.
The object and interpretant show relationships b—m and s—p, while the representamen shows only: s—b. This relationship is not directly present in either object or interpretant. The lines of association cross between representamen and interpretant, indicating that to make sense of the metaphor, we must untangle the parallelism using sociocultural knowledge.
Notice that the metaphor requires all three elements — sign, object, and interpretant — to be specified. For the sign to be interpreted correctly the parallelism to be found in the object must also structure the interpretant. But see how the representamen only has two relates to the object's four, and this missing information must somehow be restored. This is the role of the interpretant, and this is the meaning of the crossed lines between representamen and interpretant: something has to happen there, a process that I will detail below. This idea is elucidated by our utterance (b), an oft-cited example of a metaphor.
First, notice that from a syntactic point of view (that is, just looking at the structure of the two sentences), our linguistic examples of diagram and metaphor are identical, both being of the form: Noun Phrase 1 = Noun Phrase 2. However, as described earlier, (a) is a simple diagram, with a member-class relationship that can be understood and taken literally. Meanwhile (b) cannot be "taken literally" and still retain the intended metaphorical meaning: if we take it literally, claiming that this surgeon does not treat his patients like meat but really is, perhaps as a side-job, and necessarily unrelated to their role as a doctor, a butcher, our utterance becomes a diagram and not a metaphor. Notice again that it is the object that determines the representamen.
But back to (b) as metaphor. What (b) is doing is establishing a parallel (recall Peirce's definition above, which hinges on parallelism) between a "base domain" of butchers and the "target domain" consisting of the particular surgeon in question. In the base domain figures the class of butchers, who, by definition, cut up meat, saw bones, and throw big hunks of flesh around, perhaps with brute force and little finesse. The target domain, the domain that the speaker is trying to characterize or pass judgment upon, probably contains an incompetent surgeon who operates with little finesse or precision and treats her patients like hunks of meat and bone. Our surgeon is not literally a butcher; they are not in a member-class relationship with the class of butchers. Rather, there is a parallel, a resemblance, a common quality about the way the surgeon treats their patients and the way a butcher treats their meat.
This is what is represented in Fig. 4.4, with b and m representing the classes associated by inclusion in the base domain and s and p representing the surgeon and patient in the target domain. But note how impoverished the representamen is! The parallelism is quite evident in the object and interpretant, but the representamen is quite underspecified with respect to the notions of "meat" and "patient" — and "incompetence" and "lack of finesse" — all of which have to be inferred from pretty impoverished data, our representamen.
Something is obviously missing from the representamen, and these blanks must be filled in by sociocultural knowledge, like how you and I know that surgeons are supposed to be precise and competent while the work of butchers is messy and involves brute force. Notice how the same metaphor can participate in different evaluative judgments depending on the socioculture in question.
An example from fiction can be found in Jacqueline Carey's Kushiel's Dart. At one point, the heroine and another woman discuss the sexual prowess of a man, and agree that he "makes love as if he's hunting a boar" (p. 152). Later, the comparison is repeated in a different context, where a man from a different culture asks the heroine about his performance in the bedroom:
I remembered what Cecilie had said about Childric d'Essoms. "My lord makes love as if he is hunting a boar," I said; it was not as much of an insult to a Skaldi as it would be to a D'Angeline. "It is a heroic act, but not necessarily pleasing to women." (p. 398)
The same metaphor, in two different cultures, is evaluated differently. When the description is attributed to Childric d'Essoms, it is within a culture that values finesse in the arts of love, and is seen as somewhat of an insult. In the second context, however, the heroine addresses a recipient from a culture which values masculinity, strength, and skills in hunt and battle, and so the same description is much less insulting.
Notice again how much sociocultural knowledge we need to fill the gap!
Another popular example of a linguistic metaphor is "lawyers are snakes". Lawyers do not share simple properties like scales or forked tongues with snakes, nor would it make sense for lawyers to literally be a subclass of snake, so we cannot take the expression literally. Rather, our sociocultural knowledge of lawyers and snakes fills in the gaps to tell us that what we mean is that lawyers are slippery, conniving, and related the Devil (which are themselves mostly qualities we associate with snakes only by way of socioculture). Another linguistic metaphor, this time of a different syntactic form, is "a mountain of a man", where we must draw on our sociocultural understanding of mountains to infer that the man in question is probably tall, sturdy, and durable.
Much of advertising is predicated on metaphors. Jappy uses the example of a print ad featuring a woman victorious and relaxed, with great makeup, in a boxing ring with her vanquished male opponent behind her (search for "Plate 3"). To interpret this ad, Jappy says we must draw upon our sociocultral knowledge of mens' and womens' roles, boxing matches, heteronormative male-female relationships, and the role of makeup and beauty in the social power of a woman to come to the conclusion that it is because this woman has chosen the featured brand of makeup that she has won this particular round of the battle of the sexes.
One of my first projects was to write a print ad for a potato.
The National Potato Board needed to replace its current ad, which featured a potato covered in thick, green, latex paint and the headline "What must we do to make you realize we're a vegetable?" This was when they were trying to reverse the perception that potatoes were just junk food.
The new strategy was all about speed. Microwave ovens were relatively new, so speed was exciting news. And potatoes were known to be slow. So I did an ad that featured a potato in a wind tunnel, like a car. The headline was "Aerodynamically designed for speed." (p. 57)
As part of the body copy of the ad itself, Augusten included a fast and easy microwave potato recipe.
Notice that both versions of the ad, Augusten's and the version he replaced, draw on cultural knowledge that is not available from the ad itself. If we did not know that vegetables are, as a rule, green, then covering the potato with green paint would make no sense to us. Likewise, if there was no common perception of potatoes as junk food (and the underlying idea that junk food is unhealthy), and of vegetables as healthy food, then the attempt to portray the potato as (green=)vegetable would make no sense either. A similar logic applies to the second ad: wind tunnels are known to be related to speed, aerodynamics, and fast cars, while potatoes are, as Augusten says, known to be slow. Putting a potato in a wind tunnel would make no sense to us without these associations, and the general cultural discourse about convenience, speed, and modernity.
Likewise, the print ad in Fig. 4.5 calls upon understandings of sex, the game Twister, and the role of erectile dysfunction medication (note that while this is an ad for Viagra, the name of the pill itself is nowhere pictured in the ad — all that is shown is a little blue pill juxtaposed with the photo of the Twister-bed).
Figure 4.5: A print ad for Viagra featuring a bed patterned like the game Twister. Agency: Ogilvy & Mather, Mexico City, Mexico; Creative Directors: Marco Colín, José Montalvo; Art Directors: Alejandro Guadarrama, Carlos Oxté, Aurora Morfin; Copywriters: Carlos Oxté, Alejandro Guadarrama; Photographer: Flavio Bizzarri. (source
We need all these things to understand that the ad means that Viagra will restore fun to the bedroom.
Here's another ad that partakes of several modes of iconicity:
Figure 4.6: A print ad for Matrix FM Online Radio featuring a pair of headphones arranged to resemble a winking smiley face, with the copy "Enjoy from ear to ear". Advertising Agency: Nova MCP, Assis, Brazil; Creative Director/Copywriter: Eddie Silva; Art Directors: José Carricondo, Vinícius Gonçalves; Photographer: Vinícius Gonçalves. (source
This ad is pretty complex! It has both symbolic and iconic elements; both text and pictures. The literal image is a photo of a pair of headphones. However, the headphones are arranged to resemble — share qualities with — a winking smiley face (Fig 4.7).
Figure 4.7: HTC emoji of a winking smiley face. (source)
The winking face smiley or emoji are themselves images of a human face smiling and winking. It's yellow, as no human face is, but through symbolism we have come to associate yellow with the default skin colour of emoji — sort of. So we have an icon of an icon here. What is it supposed to represent? Well, from sociocultural knowledge and personal experience we know that the facial expression this ad is referencing, two degrees removed, is associated with enjoyment and pleasure. The tagline, "Enjoy from ear to ear," references the fact that headphones go on the ears. So this has something to do with enjoying and listening. The ad provides the logo of the station it's advertising, and from cultural associations with logos, we know that the ad is saying "listen to the station this is a logo of, and you will enjoy it."
These visual examples help us understand a very important feature of metaphors: the representamen is called upon to represent an object far more complex than itself. In the process, information is necessarily lost, and we ourselves have to restore it using our cultural knowledge. It is only thus that a potato in a wind tunnel can stand for the complicated idea that potatoes are fast and modern.
Now, remember our syntactic comparison of examples (a) and (b). One consequence of the impoverishment of the representamen is that diagrams and metaphors are often formally identical (meaning the representamen has the same shape or structure), even though their semiotic differences are vast and socioculturally important. The semiotic work that is done by each one is quite different. In claiming something is a diagram, we claim that the connection is direct, immediate, and already existent in the object (a Secondness). But when I make a metaphor, the relationship in the representamen is not directly available in the object, but must be inferred from sociocultural knowledge. In a metaphor, something is "missing" from the representamen, and it is the power of metaphor to play upon these missing elements. In a diagram, everything is already "there", the relationship that exists in the representamen is claimed to exist in the object.
So this is why metaphor is so powerful and interesting! Consider the problem of metaphor in translation. Does the other language have the capability to translate the words literally and come up with the same meaning? If not, how do you translate the meaning of the metaphor without using the same words?
A few miscellaneous notes on icons and semiotics and socioculture! First, our entire legal system is envisioned as an icon — specifically, a diagram — of our society. The body of laws is supposed to define relationships and actions that occur in society. I go into this in greater detail in my thesis.
Second, there is a process called "naturalization" in anthropology and linguistics. An example of naturalization is when a certain way of speaking, say, an accent, is taken to say something about the speaker. How this works is a process by which dicent symbols (rhematic symbol + rhematic index) come to be read as dicent indices (natural, causal connections) and not symbols (arbitrary, sociocultural connections). An accent is a rhematic index of the speaker by contiguity. The accent is also a rhematic symbol: it gives us an idea of the speaker because of socioculturally-determined notions about the type /speakers with this accent/. The accent is thus a dicent symbol — a proposition — about the speaker: it picks out the speaker and tells us something about them. Recall our basic example of a dicent symbol: combining the rhemes "____ is black" and "that cat" to make the proposition "that cat is black". Likewise, a British accent combines a rhematic index, "this speaker" (as any utterance indexes its speaker) with, say, the rheme "____ is British" (though other sociocultural associations may be substituted or added, such as — from certain USA perspectives — "is stuck-up", "is upper-class", "is foreign", etc.). Note that this proposition, just like the one about the cat, can be false: the speaker could be putting on a British accent and may not actually be from Britain.
Many symbols of this kind come to be read as dicent indices (remember our discussion about body language and how dicent indices can be faked as symbols). In this case, the inference is: this speaker talks this way because they are from a certain place, just as a weathervane points east because the wind is from the west. This interpretation implies a causal, rather than sociocultural, connection between speaker and accent. This is naturalization: interpreting a socioculturally-based, arbitrary connection as one of natural connection.
So!
This is why I think Peircian semiotics are a useful way to analyze writing: you can really dig into how the writer conveys meaning, what knowledge they are drawing on, how the writing works. Of course, I also come with a detailed knowledge of English language norms to analyse the writing on an editing level. But I hope this series has been an interesting look into how I think about language and writing!
___________________________
One last review:
Image
A type of icon that shares simple qualities with its object.
Diagram
An icon that shares an analogous relationship in the representamen with the same relationship in its object.
Metaphor
A type of icon that represents a relationship in the object by showing a relationship between something else in the representamen. To reconstruct the relationships and obtain the interpretant, sociocultural knowledge is required.
Thanks for reading! Next post, I promise, ACTUAL BOOKS AND WRITING.
Notes
Some of the examples and the form of all the sign pictures in this section are taken from Jappy, although I made my own versions of the figures. (back to text)
For example:
Turino (1999) (yes, again, he applies Peircian semiotics to music)
Mertz, E. (1996). Recontextualization as Socialization: Text and Pragmatics in the Law School Classroom. Natural Histories of Discourse. M. Silverstein and G. Urban. Chicago, University of Chicago Press: 229-249.
The onomatopoeia for the sound a cat makes (mewing, not purring), is one of the most similar representations across languages.
The word for "mother" is also quite similar across most languages. My favourite theory as to why (aside from most of these languages having Indo-European roots, but not all of them do) is that it has to do with the shape of the mouth babies make when nursing: make that, add a vowel, and you get "mother" is a lot of languages.
The following example, and its formulation and discussion, are in large part owed to Jappy, so much of the terminology and structure — though little of the exact phrasing — of the following must be attributed to him. (back to text)
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
justira posting in ![[community profile]](https://www.dreamwidth.org/img/silk/identity/community.png){kind=small}
ladybusiness
