Clipart Images as Common Sense Categories

(forthcoming in Visual Communication)

As PDF with color


Clipart is the visual representation of commonsense categories and stereotypes—of social representations, in the sense of Moscovici. As indexed and displayed on line for downloading on Microsoft's Design Gallery and Art Today's, the clips exhibit strong threads of continuity with earlier popular illustration and rhetorical manuals of gesture. At the same time, they also register differences in the social representations of different eras, as well as divisions and complexities in the representation of controversial categories such as sexual harassment. Though it is questionable how and in what way the theories popular in an era can be critiqued by citing the clipart of that era, these large archives provide powerful new tools for seeing, displaying, and investigating social representations-more powerful and useful than directly accessing Google's hundreds of millions of on-line "images."

Figure 1
(©2001 Hillustration Studios)

From the point of view of high culture (and professional illustrators), clipart is junk: it may be used to illustrate, but it is almost never clarifying in the way a good diagram can be. Because it is pre-made, it can only be roughly right for the current meaning, and it has a great penchant for illustrating the commonsense categories and stereotypes of common understanding. Pasting it in is much cheaper and easier than hiring an illustrator to make original, custom visuals. And, as Figure 1 suggests, it is something people with fine taste (and servants) contemptuously dismiss. All of these attributes are reasons why it has found no place in academic publishing, flourish though it may in desktop publishing. Indeed, many academics are unaware of the huge archives of clipart on sale and available for free from their desktop computers. These same attributes, however, make clipart a very useful site for investigating the representations of the everyday in non-high brow, non-academic culture. These representations portray the world as common and familiar; clipart aims to be recognized as bits of received, conventional wisdom, commonsense categories, or social representations. Of these several roughly equivalent terms, I will argue that Serge Moscovici's term social representations best captures the purpose and function of clipart and indeed, that studying clipart is a tool for exploring social representations.

The various scholarly approaches to commonsense knowledge (or beliefs) over the last few decades fall into two camps: those who describe this knowledge as sets of propositions, maxims, scripts, or other semantic structures (van Dijk, Billig, Schank and Abelson) and those who describe it in terms of images and prototypes (Moscovici, Rosch). To a certain degree, this difference can be attributed to a difference in aims: the first group is interested in commonsense argument and prejudice; the latter emphasizes categories and attitudes--the way the world is seen (pre-argumentatively, as it were).1

Both schools understand common sense to be shared, taken-for-granted views of the world. These views are somewhat abstract, generic views of typical things, events, and relations (Schank and Abelson discuss the "coffee shop track" of the "restaurant script;" Moscovici speaks of the image of a "Frenchman" and of a "neurotic") with unspecified or defaulted specifications of various details that are filled in (instantiated) as one experiences particular individuals and situations. Moscovici describes the social representation (or the visual part of it) of a Frenchman as an undersized person wearing a beret and carrying a long loaf of bread (Phenomenon, 43); contemporary on-line clipart archives furnish numerous examples of that image, but also popular as markers of Frenchness are cigarettes, wine, horizontally striped (sailor's) shirts, and moustaches. His discussion of "neurotic" is less visual, using it to illustrate the "decay" or sinking of a term of scientific culture into a social representation (of common sense).2 Gerard Duveen, in introducing the concept of social representation, observes that those who grew up during the Cold War would place Prague to the East of Vienna on a map of Europe because of the social representation of East Bloc countries lying to the east of the West Bloc (Duveen, 1). The general thrust of my argument is that clipart archives give us a way of investigating the visual part of social representations.

Both conceptualists and imagists agree that the whole body of common sense is not logically consistent or coherent, nor critically reflected upon (hence its taken-for-granted quality) and is a mixture of traditional ways of thinking, scientific theories, media representations, personal experience, and, as we say, urban folklore. There are of course belief/images with these qualities that are not held in common but only shared by a certain group and are points of contention with other groups (family, couple). These van Dijk calls ideologies (van Dijk, 8). And we will add another status-alternative views about which the culture as a whole has reached no decision (what is a hacker? What relation to a geek?) and may be in the process of change.

The parenthetic examples here indicate that my first interest is social categories ("views") rather than arguing and reasoning, but there is no reason to assume that one mode of representation must be chosen over the other. Moscovici in fact speaks of social representations as "a network of concepts and images tied together in various ways" (Notes, 222). We maintain dual, cross-referenced systems of images and verbal representations rather like the clipart archives that have now become available online. These archives are databases with images categorized in various ways using multiple key words. They thus can be used as word-to-image dictionaries ("Computer-show me images of geeks!"), especially because clipart images depict the familiar and established views of things. We might speak of the keyword hacker, for example, as polysemous between a cluster of images of slobby computer fanatics and a cluster of malicious, destructive thieves (with some overlap between those images in the first cluster and those associated with the keyword geek). Since the word-image mappings in two archives are many-to-many, these archives have the form of word-image networks, and since the numbers of their images are in the millions, these networks are very large. They are not all equally illuminating, however, for the subtle intelligence of a lexicographer is needed to make good matches between image and word. The purpose of this article is not to review word-image archives, however, but to use them to argue first, that clipart means to and does represent the categories of common sense and second, that clipart can be used in a critique of common sense.

1. Images of Categories

The clipart that is our concern can be found in two large archives with keywords: Microsoft's Design Gallery Live and the archive at ArtToday (clipart. com). In addition, I will refer to a third archive, that of Google, which is not cross-referenced in the fashion of the other two, but does provide the largest body of images associated with common terms found on line. (It has, for example, 1140 "images" associated with neurotic.) These archives contain photographic clips as well as the more traditional and central black and white drawings and vector graphic drawings (these most often in color). In fact, ArtToday includes clips drawn from the history of print illustration over the last centuries. Drawing from popular books (Compton's Picture Encyclopedia), classics of childhood (The Book House for Children, Uncle Arthur's Bedtime Stories), illustrated magazines (Scribner's Illustrated, Punch, The New Yorker, and Dover collections of "vintage" and period clipart, the history of popular imagery provides a substantial portion of ArtToday's 2.5 million clips. The clips vary in style of representation as well as medium: a sizeable proportion of them are cartoon-like in exaggeration of various features and may even attempt humor. Some do not represent the category directly, but via metaphor and some use of diagrammatic conventions (arrows, lines of radiance). The majority, and what I will consider the central body of clipart representations, simply try to represent the thing as it looks. We will sample these archives for their representations of certain terms for human roles, experiences and activities (teacher, reading, and thinking) but also for some more abstract concepts (sexual harassment, discrimination, stereotype).

If the object is to present the thing as it looks, one would suppose that photographs would be ideal and, in digital media at least, would replace drawn clips. Although the archives have been adding photo clips in recent years, they still do not amount to more than about a tenth of the comparable body of clips for a particular category. This is surprising on the face of it, since photographs have the advantage of realism.3 Modern digital displays and printers can display photographic images as effortlessly as line drawings or silhouettes. If recognition of the familiar is of the essence, why not a photo, with its familiar stylizations of the phenomenal world?

One problem with photos is that they are so grounded in a unique scene with unique people and things in it at a unique day and time that it does not do the work of categorization for us-it leaves it to us to pick out the respects in which the individuals and setting are instances of the categories.

Figure 2

Figure 2, for example, is a photoclip with a keyword teacher. It uses standard zone-of-focus blurring to pick out the important figures from the background (except for the too-salient backpack), but it invites us in to think about the lives of the individuals (What is troubling the younger woman? Why is the teacher concerned? Who else is watching?) rather than to say, "Yes, that is a teacher teaching." Figure 2 is not only an image of a teacher: it has in fact nine keywords in ArtToday: teacher, student, talking, classroom, conversation, embrace, education, school, and backpack.4 Insofar as a photoclip appears to be taken "from life" it draws us in to those subjects and their lives and away from the category to be depicted. To put the point grammatically, we are trying to depict a common noun, but photography pushes us in the direction of proper nouns.

This conclusion would have come as no surprise to Otto Neurath, whose ISOTYPE geometric pictographs are one of the major tributaries feeding the modern body of clipart. Neurath, who was one of the Vienna Circle of logical positivists, developed the International System of Typographic Picture Education (ISOTYPE) in the period between the World Wars, hoping that simplified and standardized visual symbols could advance both scientific rationalism and communication among people, especially those with limited formal education, along with a progressive, democratic socialist political agenda. Neurath advocated the silhouette style of representation because it abstracts from particularities that have no part in the proper understanding of a concept and because it represses the look of free-hand drawing in favor of regular principles of projection from thing to sign.

Some of these particularities, he held, are the attributes of commonsense stereotypes, where the "common" sense is anything but common across cultures and is the antagonist of scientific rationalism. One might in fact question whether the teacher figure with the pointer and globe is sufficiently stripped of inessential details to meet Neurath's criterion, since the teacher is marked as female by her slightly flared skirt (essential for women's rooms, but not classrooms) and is depicted with a pointer and a globe, as if geography were the only subject. In fact, of course, the pointer and globe are conventionalized markers of teachers in clipart and do not represent anything essential to the concept of teacher, at least not by principles of geometric projection. ISOTYPE style was well suited to the very limited graphic resources available in the early days of desktop printing, and there are still 76 clips in this style (Style 341) in the Digital Gallery Live archive despite the availability of high resolution and millions of colors. Black silhouette style (Style 30-lacking the geometrical abstraction) is found in over 1000 clips in the archive.

Today Neurath's social program seems touchingly retro, and the goal of universal scientific rationalism is in retreat before the advocates of localization, but the impact of ISOTYPE endures in public signage and the display of statistical information, and in clipart, where it has become a challenging stylization for the illustrator. ArtToday has a large number of them (gathered from collections by Dover Publications), including a series of men committing suicide by gun, poison, gas, hanging, leaping.5

Figure 4

Figure 4 is quite disconcerting, and probably humorously intended, since we associate the ISOTYPE style with regulative signage—"this is what to do here"—and there is no time or place proper for committing suicide. It is remarkably cold-blooded, suitable mainly for the pictographic bar charts ("Numbers of Death by Cause") that Neurath was very fond of.

If photographs anchor down the concrete, detailed end of the representational scale, ISOTYPE anchors the abstraction end. A diverse range of intermediate styles have been developed, many of them using category markers such as the pointer and globe in Figure 5.6 These markers (along with the "apple for the teacher") occur quite frequently in the clips for teachers and virtually never in the photos (once out of 150 photoclips for teacher in ArtToday, and that one is miscataloged as clipart). These markers are another reason not to think of clipart as drawn "from life."

There is another branch of clipart that diverges from simple sketching of "what we see"--namely the branch identified in Design Gallery Live with the keyword metaphor. There is a particularly fine example in ArtToday illustrating computer addiction by means of a net that comes out of the monitor, surrounds the addict, and draws him back into it. Some of what DGL calls metaphors are what ArtToday calls symbolism and clichés-literalizations of sayings such as "climb the ladder to success," "climb on the backs of others," "reach for a pie in the sky," "jump through hoops," "wear many hats," "print money," "stretch a dollar," and so on. Both main archives have plenty of these, which are strongly motivational, as they say in business. These are almost word-for-word illustrations of the sayings and make no visual sense without the sayings-they are like visual puzzles waiting to trigger the verbal cliché and so solve.

Figure 6

Figure 6 has keywords annoyed, frusterated, frrazzled, and cliché, but fortunately it also has the keys get, out, of, my, hair. Thus the primary form of the categories in these cases would seem to be verbal rather than visual. Others of the metaphorical group are mixtures of visual representation and diagrammatic markings such as arrows and activation beams. Yet others are built on simple substitutions of light bulbs for ideas (or thinking, if activated) and flowers for "growing." They take a little figuring out, which makes them useful as attention-getters but makes them less likely candidates for basic representations of categories. Finally, there is the group ArtToday calls abstract. The clips in this group strive not to look like other representations of the things but to offer another take on them, or to move toward pure form and color, or to uninterpretable enigmas. None of these would be suitable depictions of basic social categories.

Figure 7

These then are some of the resources of clipart that may be particularly useful when inventing clips for new social representations and some of the harder cases than teacher. How can clipart represent complicated and contentious things like sexual harassment and sexual discrimination, which are only slowly being integrated into common sense? How could it convey stereotype, which is to represent itself? The Design Gallery Live cannot help us, for it does not recognise any of these as keywords. ArtToday gives us five clips of sexual harassment and several from older illustrations. In addition, it gives two examples of sexual discrimination, one a clip (Figure 7) and one from older illustration (a political cartoon by Alice Beach Winter, c. 1912). Figure 7 makes visual the term glass ceiling and is an example of metaphor, visual and verbal. The older illustration shows a girl reading a "Boys Only" sign in a butcher shop window-which makes its point, though it has to use words to do it. It only illustrates discrimination at the hiring level, of course, but it is still not bad for a 1912 socialist political cartoon.

Figure 8

With sexual harassment the matter is more complicated. The older illustrations, which involve the man's groping the body and breasts of the woman, are largely unproblematic (though technically they illustrate the crime of sexual assault). Three of the four clips, however, are also found among the 167 clips linked with flirt, even Figure 8, where the man's attentions appear clearly unwelcome. A few others in the flirt set are also candidates for harassment, though not so classified by the archivist. Even worse, some of the flirt clips also illustrate sexual assault, including Figure 8. And so one's hopes to find a set of clear illustrations for a brochure on sexual harassment, or a poster announcing a training session, are disappointed, but this zone of overlap and uncertainty also reflects the current state of things in the commonsense view of the matter.

We can find photo images from training manuals and presentations in Google-Images, the archive with no less than 1430 images returned for the search term sexual harassment. Figure 9 is a grayscale reduction of a screen capture of the 20 images and illustrates many of the main points of difference of the Google image collecting/cataloging system.

Figure 9

The Google web-crawler apparently assigns this search term to pages it comes across that contain sexual harassment as an image title, caption, or anywhere on the page, so the image may not be illustrating the category of sexual harassment. The first row has one clip depicting fanny-pinching at the office copier machine (a case, again, of sexual assault, though stolen rather than compelled) and the second has three posed photos from various training manuals (one of which is not such a clear case). There are one or two others, but most of the rest are posters, pie-charts, arcane cartoons, and depictions of scenes possibly related to sexual harassment.

It is even harder to sort through the images to get a few clear, prototypical images of sexual harassment. There are about 40 other images, many of them "practical joke" forms of corporate humor indicating resistance in offices to policies and training sessions and thus resistance on the part of (presumably male) office workers to accepting s.h. as a taken-for-granted category of common sense. Interestingly, though, these 40 images are linked to the less standard spelling "sexual harrassment."

The difference between Google and the others is that the data has been gathered without any sorting or selection of prototypical scenes or editorial control of the assignment of the (single) category keyword. It is in effect a data dump of what a superlexicographer might collect as input, so, in effect--"you be the lexicographer." Before accepting the challenge, however, you should consider that there are 4080 "images" for harassment (of all kinds) and 327,000 images forteacher. At present (August 2003). The point is that Google's "images" are not a profile of the social representations/categories of common sense in question, though they provide a great deal of indirect data about what they might be.7 But that is what the clips in the archives set out to be and what people think they are when they use them.

Further, if we look just at the images from the training manuals, we see again the problem of uniqueness with photographic clips: even though the figures are most likely posing models, we want to know more about the relationship of the individuals depicted and this is a unique one with its own history and multiple perspectives. In contrast, the cartoon exaggeration of posture and expression in Figure 8 help to highlight the criteria that distinguish sexual harassment from flirting (or warm regard).

Figure 10

Dictionaries define stereotype in terms that sound like learned culture describing commonsense understanding. Merriam Webster's Online Dictionary says "2: something conforming to a fixed or general pattern; especially: a standardized mental picture that is held in common by members of a group and that represents an oversimplified opinion, prejudiced attitude, or uncritical judgment." Using the word as a query term in ArtToday yields eight images, four of them clipart. Two represent animal behavior stereotypes (a mouse happily eating cheese and an ostrich burying its head) and two depict American Indians: one is Figure 10 and another is of a brave rampaging with a tomahawk. These are stereotypes known to be such; they are known to be the products of movies, "humorous" caricature, and sports publicity, the heaviest users of "oversimplified opinion." These clips, insofar as they represent commonsense categories, indicate that common sense has some capacity to reflect on itself, some way of sorting through images critically. To put it another way, to have common sense is to know what the stereotypes are.

Figure 11

The stereotyped Indian chief appears in another cartoon clip (Figure 11) where a further stereotype of the Native American "peace" pipe as filled with marijuana is invoked, the clip being the unique link to the term stoned. He is without a doubt a stereotype, but he is good for a knowing laugh at the paleface tourist's expense.

Figure 12

Common sense also has the capacity to change: images that at one time might have seemed bits of genial good humor now seem condescending relics of an oppressive system-they have become prejudicial stereotypes. "Aunt Jemima®" is an example of an image that sold a lot of pancake mix before it was assailed for echoing the Mammy stereotype from blackface minstrel show and plantation imagery; refurbished, however, more youthful and sans bandana, she is still in service selling pancakes.8 Figure 12 is linked in ArtToday with the keyword stereotype.

2. Clipart and Critique

Some people think of clipart as one of the resources for desktop publishing that rather suddenly appeared sui generis with advent of personal computers. Some of that sense appears in the following comment by Richard Lanham, even as he tries to overcome it by connecting clipart to common sense and both of them to traditional education in expressive gesture:

Digital expression has resurrected the world of proverbial wisdom, but through vast databanks of icons rather than words. We buy what are, in effect, catalogs representing commonplace situations and appropriate responses to them: faces, hand gestures, signage of all sorts.... The traditional dependence on commonplaces in rhetorical education has been transmuted from word to image. (Lanham, 1993: 37)

Lanham is right, as we shall see, to point to the tradition of rhetorical education as a source of clipart social representations, but it is not evident why he speaks of a resurrection of the world of proverbial wisdom. Has he not read a magazine, a newspaper, a catalog, seen a television show, a movie? Commonsense knowledge as we are using it is equivalent to proverbial wisdom: both terms connote continuity with the past that strongly constrains innovation and novelty. A crucial part of Moscovici's theory of social representations is just this continuity. He says,

All the systems of classification, all the images and all the descriptions which circulate within a society, even the scientific ones, imply a link with previous systems and images, a stratification in the collective memory and a reproduction in the language, which invariably reflects past knowledge, and which breaks the bounds of current information (Moscovici, 24).

To see just how deep and consistent this tradition of representation is, we can sample some of the handbooks of rhetorical gestures that were indeed part of rhetorical education for centuries and well into the nineteenth century.

Figure 13

The first pose is copied from Gilbert Austin's Chironomia, or, a Treatise on Rhetorical Delivery (1806). Concerning Gesture 87, Austin says: "Deliberation on ordinary subjects holds the chin, and sets the arms akimbo" (Gilbert Austin, 1944: 490). The second pose is from Century Magazine, Vol. 7 (before 1880). All of the last three can be found in the ArtToday clip archive under keywords contemplation or pondering.

Figure 14

The first part of Figure 14 is drawn from John Bulwer's Chironomia, or the art of manual rhetoric (1644), a very popular, many times reprinted work. Of this gesture, captioned in the fashion of emblem books with the Latin tag Inventione laboro (I work in discovery), Bulwer says, "The finger in the mouth gnawn and sucked is a gesture of serious and deep meditation, repentance, envy, anger, and threatened revenge" (John Bulwer, 1974: 121) and he cites numerous lines from the Roman poets about biting nails while writing and as a sign of envy and the like.9 The second appears to be a Louis Rhead illustration from 1923; the third a Jill Elgin illustration from 1955; the last from Design Gallery Live.

Figure 15

Figure 15 also comes from Bulwer and is captioned Sollicite cogito (I compel by repeated requests): "To rub or scratch the head with the hand is their natural gesture who are in anguish or trouble of mind; for commonly when we are in doubt and uncertain what to do, we musing scratch our head." (John Bulwer, 1974: 72) The little guy behind the eight ball comes from the 1920s, and the third is from Design Gallery Live. All of these are the lightest of samplings: the archives contain several other clips/illustrations (at least) for each representational style and keyword. But regardless of style or century, the marking gesture is repeated over and over-a purely visual element that is as entrenched as any bit of word meaning.

Style and era do introduce inflections in the basic language, however, and period clips readily evoke a whole cultural milieu in which they flourished. If one wants to argue that a certain idea or understanding of one period is faulty, its own images can be used against it. Clipart images lend themselves to this deployment, because they suggest that the idea under attack was in fact a widely held received opinion.

An example of this mode of arguing can be found in "Blinded by the Letter: Why Are We Using Literacy as a Metaphor for Everything Else?" by Anne Wysocki and Johndan Johnson-Eilola (1999). This piece argues that "computer literacy" is not a good term for describing the relations we hope people will develop with computers and offers a few suggestions about other ways to think of those relations. It begins by characterizing two "bundles of stories" that are widely believed, one that literacy is a necessary and sufficient key to improve anyone's life, and the other that the book is essential to life as we know it:

To the book, then, the writers we have quoted attribute our sense of self, our memories, our possibilities, the specific linear forms of analysis we use, our attitude towards knowledge, our belief in the authority of certain kinds of knowledge, our sense of the world. (Wysocki and Johnson-Eilola, p. 359).

Some of the authors they have quoted include Marshall McLuhan and Walter J. Ong, and their discussion suggests that these views are old-fashioned, noting that they have been severely criticized by more recent scholars.

This first, negative phase of the argument is illustrated by six bits of clipart. These clips are the sort that one might find in schools and libraries advertising the importance, pleasure, and wholesomeness of reading. Or might have found: Wysocki and Johnson-Eilola observe in an end note that in all the images in the collection,

all the people are white and clearly middle class; there are many illustrations of women and children holding books; if men have printed matter in their hands, it is account books or newspapers-unless they are shown reading to their families (as one illustration here shows). (Wysocki and Johnson-Eilola, 1999, 368)

They do not spell out the implications, which I assume are that the collection is skewed in the direction of the racial and gender stereotypes that prevailed when the clipart was made. (Men do serious reading to understand and master the world; women read to indulge escapist fantasies; see Long, 1992.) They use the clipart to illustrate the stories of literacy they are attacking, again assuming that the stories were in full flower at the time the clips were published. Even if the bad stories didn't cause the biased depiction of literacy, they are damagingly tainted by the racism, sexism, and classism endemic in the era and reflected in the skewed representations of the scene of reading. Here is an example of how method of argument works.

Figure 16

Figure 16 is placed immediately following text expounding upon the supremacy of the book as proclaimed by conquistador priests and recently Sven Birkerts. The text alongside the image says: "Birkerts' words call to our minds Habermas, who wrote that a necessary (but not sufficient) step in the development of a critical public in the 18th century was that men read to themselves: in the privacy of their reading they developed a sense of individuated self, a self that would hold a position in the public sphere" (Wysocki and Johnson-Eilola, 1999, 358). The clip presumably represents the young man looking out from his book-nourished interiority toward his future in public life (the hurricane lamp behind him perhaps evoking the domes of public buildings). His vest and haircut place him as middle-to-upper class, circa 1930-1960. This actually suits him as an exemplar, though distinctly American and a century or two belated, of Habermas's inventors of the public sphere, which was, his subtitle reminds us, "A Category of Bourgeois Society" (Jurgen Habermas, 1962/1991). As we look at the lad's image today, we see him as one of the privileged ones of his time, well furnished, well dressed and groomed, handsome, and engaged in his studies as he should be. He exemplifies the "clean teen" image that Steven Heller says was promoted in the 1960s to combat the "juvenile delinquent rock-and-roller" image (Heller, 89-93).

Figure 17

Little or none of that is conveyed by Figure 17, a contemporary clip. Here clearly the book is not the only thing nourishing the lad's interiority. This contemporary version of reading makes no link with forming the imagination and vision or with class membership and promise for a position in the (bygone) public sphere. In fact, it doesn't promote anything. A similar contrast could be made for another of Wysocki and Johnson-Eilola's 60's clips of a father reading to his two pretty white, well-dressed and groomed model children attentively sitting on each side of his lap, where the accompanying text is the heavily critical comments of Brian Street: "[T]hese grandiose claims for 'academic literacy.' he writes, "are merely those of a small elite attempting to maintain positions of power and influence by attributing universality and neutrality to their own cultural conventions" (Street, 224, cited in Wysocki and Johnson-Eilola, 354). Here the equating of the 1960's clipart and the old wrong "story" about literacy is complete. Unfortunately for Wysocki and Johnson-Eilola, Figures 18 and 19 indicate that their resistance to promoting "computer literacy" in these ways is a lost cause:

Figure 18

Figure 19

If the argument about skew (or worse, elision or suppression) is to be made from a collection of clipart, there should be some reason to think that the collection is comprehensive or representative. Unfortunately, the collection under discussion is as they say small and perhaps largely the work of the Harry J. Volk Studio alone. If a major collection such as the Design Gallery Live or ArtToday were to be made for the clipart of say each decade, we could characterize the special qualities and exclusions of different decades and we would have a basis for saying that the clipart representations of the era were skewed or blind to some things represented in other eras. We would have an excellent way to show how the culture of an era represented itself to itself. But at the current time, the only era so supplied with such databases is the present, and these unprecedentedly huge collections differ in their profiles of contemporary social representations.

Comparing the two large indexed, on-line archives, it is quickly apparent that ArtToday holdings are not only much larger in the categories we have been studying (teacher, reading, thinking all have about three to four times as many clips in ArtToday as in Design Gallery Live) but has clips for many categories where DGL has none, or sometimes one. The greatest differences are for slang and "unmentionable" topics (the first figure is number of ArtToday clips, the second Design Gallery Live):

Table 1: Differences of Clipart archives in sensitive domains

SLANG: nerd (54, 0), hacker (65, 2), buzz [intoxicated] (6, 0), horny (2, 0),wimp (4, 0)

POOR: beggar (21, 2), bag lady (2, 0), homeless (49, 0), poor (61,7)

DRUG USE: snorting (1, 0), shooting (2, 0), stoned (1, 0)

SEXUALITY: gay (63, 4),topless (13, 0), dominatrix (2, 0), sexy (33, 0)

LOSS OF CONTROL: insane (31, 0), crazy (69, 4), lunatic (10, 0), loony (6,0), nuts (6, 0), screw loose (4, 0), drunk (118, 1), suicide (25,1)

NAUSEA: vomit (1, 0), puke (3, 0), nauseous [sic] (3, 0)

Summing across all categories, ArtToday outpoints DGL 618 to 14. The most significant number is zero: for nineteen of these categories, there is no clipart image in the Design Gallery Live. Hence someone looking for clips to give clues to the meanings of these terms in English would draw blanks.

If we think of these archives as the collected social representations of particular groups, then the DGL group do not get down to the street very often, or at least find no need to communicate about it. The DGL title banner actually reads "Microsoft® Office Design Gallery Live," reminding us from the outset that it is oriented to life in an office, one where Word and Excel and PowerPoint are installed on every computer. Within its chosen limits, it includes people with diverse skin color and features in the clips and depicts women in all kinds of employment. It is true that in some cases one must scan a few pages to find them, but the keyword indexing makes it possible to narrow a search quite efficiently. Nothing is "buried in the back pages." Unlike ArtToday, it has no interest in displaying the clipart of former eras. In that sense, it is all contemporary; even though some of its Styles strongly evoke certain periods (e.g. Style 3 is 1950s, Style 43 is 1940s, Style 29 is 1920s-30s), they are not identified as such.

We should thus be cautious to the point of wary of skew arguments projecting what is/is not in a clipart archive onto the mentality, beliefs, and attitudes of whole cultures and whole eras. What invites this projection is the term common in common sense. If clipart represents commonsense views and categories successfully enough to be a commercial success in its time, why should it not be taken as a profile of the whole cultural mentality at that time, just as a dictionary of common and colloquial language would be? To be sure, no one has ever made such a dictionary, and these archives are not compiled with the thoroughness and method of a modern lexicographer. Compilers of these archives are not explicitly committed to inclusion of all the images currently in use and they have no standard procedure for assigning keywords to images. But we have as yet no way to classify and sort images except through keywords and verbal descriptions. In fact, it is remarkable that the two large, on-line archives are as similar as they are in keywords and as comparable. And because they are comparable, we have good grounds for saying that certain areas of common life are scantily represented in DGL, even adjusting for the much larger collection of ArtToday.

Pursuing the analogy to lexicography, we may view Google's Image database as a corpus of word-image pairings, not a sampling but rather closer to the entire body of such pairings: Google says at the time of this writing that it is searching 425,000,000 images. What it needs is an interpreting intelligence to epitomize the common-sense views and to discard a great deal of distracting detail, the vestiges of particular uses and contexts. And that is what the clipart illustrator does and why clipart provides a special, extensive but focused view of social representations.

University of Washington

© 2003 George L. Dillon


1 Confusingly, Moscovici's term social representations is adopted by van Dijk to mean “organized clusters of socially shared beliefs” (Ideology, 46) and beliefs for him are definitely propositional in nature, so Moscovici's emphasis on the visual disappears.

2 Moscovici cites an article from the 1978 International Herald Tribune speculating that the term neurotic, having been dropped as a diagnostic category of the American Psychiatric Association, would disappear, at some loss, from common usage. And indeed, there is only one "clip" in the archives for it, a 1944 cartoon by Bennett Cerf.

3 Gunther Kress and Theo van Leeuwen claim (correctly, I think) that "for us, as commonsense viewers," the 35mm camera, lenses and film sets the standard for realistic depiction of the world -Reading Images, 163. That, however, is now beginning to change to the look produced by digital cameras, which handles light quite differently.

4 Recently, several companies (Corbis, Corel, and Hemera) have introduced a sort of hybrid between a full photographic image and a clipart drawing. Called Photo Objects (and in one case trademarked as Photo-Objects®), these are photos of figures with the background masked away. That is, the image appears with no visual context behind and around it. Oddly enough, Corbis decided to make its two sample teacher photo objects "eye-catching" by mugging and distorting camera angles and by posing the models with pointers-markers of teachers from the clipart tradition. All of the objects with people have them facing the camera straight on (albeit from below sometimes), which turns them into "appeals" (Kress and van Leeuwen). To be sure, these objects are part of Corbis' Bizpresenter package of images for businessmen to use to enhance their PowerPoint presentations, so a somewhat specialized sense of humor is being appealed to. Hemera uses the same posing and shooting from high angles, at least in their sample suite.

5David Sless reports an international study from the 1980s showing that comprehension of 108 ISOTYPE symbols, 32 of them in common usage, was less than 50% for 86 of them (David Sless, 1986: 21).

6 This pictograph occurs in both of the major archives and is taken from the Dover Pictographs of People, ed. Phillip R. Rumquist.

7 Some of Google's images are the clips we have been looking at, so in that sense Google does include some direct social representations.

8 As noted, AJ is a registered trademark of the Quaker Oats Company. Every company desires its trademarked term and/or logo to become a commonsense category. There are, however, no AJ clips, and I class AJ as a popular familiar brand name. See

9 Bulwer seems to be conflating two hand-to-mouth gestures here: one of concentration, and "the thumb bite"- an insulting gesture no longer current, but still practiced in Bulwer's time and before, as in beginning of Romeo and Juliet, when the Capulet servant Sampson starts a quarrel with the Montagues with the gesture. See Armstrong and Wagner, pp. 31-32.


Armstrong, N. and Wagner, M. (2003) Field Guide to Gestures. Philadelphia: Quirk Books.

Austin, G. (1966) CHIRONOMIA, or, A Treatise on Rhetorical Delivery [1806]. M. Robb and L. Thomssen (eds). Carbondale, Ill: Southern Illinois University Press.

Billig, M. (1987) Arguing and Thinking. Cambridge: Cambridge University Press.

Bulwer, J. (1974) CHIROLOGIA: or the Natural Language of the Hand and CHIRONOMIA: or the Art of Manual Rhetoric [1644]. J.Cleary (ed.), Southern Illinois University Press.

Duveen, G. (2001) 'The Power of Ideas.' Introduction to Serge Moscovici, Social Representations. G. Duveen (ed.) New York: New York University Press.

Van Dijk, Teun A. (1998) Ideology. London, Thousand Oaks, New Delhi: Sage Publications.

Habermas, J. (1962/1991) The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society. Cambridge, MA: MIT Press.

Heller, S. (2002) The Graphic Design Reader. New York: Allworth Press.

Kress, G. and Van Leeuwen, T. (1996) Reading Images. London: Routledge.

Lanham, R. (1993) The Electronic Word . Chicago: University of Chicago Press.

Long, E. (1992) 'Textual Interpretation as Collective Action,' in J. Boyarin (ed.) The Ethnography of Reading, pp.180-211. Berkeley, CA: University of California Press.

Lupton, E. (1989) 'Reading ISOTYPE' in V. Margolin, (ed.) Design Discourse, pp.145-155. Chicago: Chicago University Press.

Moscovici, S. (1988) 'Notes Towards a Description of Social Representations.' European Journal of Social Psychology, Vol. 18: 211-250.

Moscovici, S. (2001) Social Representations. New York: New York University Press.

Rosch, E. (1977) 'Human Categorization'. in N. Warren (ed.), Studies in Cross-Cultural Psychology, vol. 1. London: Academic Press, 3-49.

Schank, R. and Abelson, R. (1977) Scripts, Plans, Goals, and Understanding. Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Sless, D.(1986) In Search of Semiotics. London: Croom Helm.

Taylor, C. (2000) posting to InfoDesign cafe ( April/58642.html).

Wysocki, A. and Johnson-Eilola, J. (1999) 'Blinded by the Letter: Why Are We Using Literacy as a Metaphor for Everything Else?' in G. Hawisher and C. Selfe (eds) Passions, Pedagogies, and 21st Century Technologies, pp. 349-368. Logan, Utah: Utah State University Press.