Photographic Inspirations: Robert Frank, The Americans (1959)

The title of Robert Frank’s The Americans knowingly places it in conversation with Evans’s American Photographs from 1938, which Frank used as an 'iconographical sourcebook' (the phrase is Tod Papageorge's). Indeed, Evans sponsored Frank’s application for the Guggenheim Fellowship he used to finance the cross-country trips on which the images were taken, and these trips sometimes followed itineraries suggested by Evans based on his travels while employed by the Farm Security Administration in the 1930s. But there are important differences as well as similarities.

Evans was interested in what he took to be a shared American vernacular culture. And for the FSA, if not for Evans, the disadvantaged human subjects of his photos were all aspirant participants in a shared American dream. They were just down on their luck, and needed a helping hand to participate fully. But by the time we get to Frank, there is no unified community to which his book's title might refer; or rather, for Frank, 'society is an association of minorities', to use a phrase John le Carre gives to his character George Smiley. The Americans therefore depicts a multiplicity of subcultures: bikers, Jews on Yom Kippur, a portrait of three trans people that is very striking, not only because such people were rarely depicted in documentary photography of this period, but because it’s one of only two images where Frank's subjects are smiling and playing to the camera. 

The Americans may not be a unified vision of America, but it is a very inclusive one, with many African Americans, as well as Latinos, possibly Native Americans (two hitchhikers driving Frank's car) and, indirectly, Asian Americans (via the photo of a 'Chinese cemetery'). One might say that Frank occupies one of several roles in relation to these various milieux or subcultures: a passer-by, a visitor, or sometimes perhaps a guest, but he's never a member, still less a participant. 

There are several examples of what I would call temporary pseudo-communities: bars, a casino, a factory, crowds at a film premiere, kids gathered round a jukebox, men crammed together along the counter of a diner (but all intent on their own eating). The political events also fall into this category, since there’s an emphasis on the hollowness of the rhetoric and the underlying realpolitik of wheeling and dealing: the grimacing or sneezing city father on a platform; the tuba player with an Adlai Stevenson badge whose instrument obscures their face so that its parping mouth comes to stand for the rally as a whole; the small group of men huddled together like mafiosi in a convention hall, oblivious to the larger crowd behind them ('sleek face earnest wheedling confiding cigarholding union boss fat as Nero and eager as Caesar in the thunderous beer crash hall leaning over to confide' as Jack Kerouac puts it in his introduction to the book).

Many pictures, however, show people alone – distracted, miserable, or just silent and self-contained. Or, if not literally alone, then psychologically so – for example, the man getting his shoes shined in a Memphis men's room. This image is particularly interesting: nominally a portrait of deference and privilege (the black man bent down at the feet of the white), but here transformed into what Kerouac calls 'the loneliest picture ever made, the urinals that women never see, the shoeshine going on in sad eternity'. It's tempting to read the hand over the man's face as a gesture born of shame, but more likely Frank just waited for a moment when the man's view of him was occluded, so he couldn't see Frank take the picture. This is also one of many photographs in the book depicting people engaged in service work (waitresses, a nurse, an elevator girl).

The larger idea of the country is still there, but it's represented primarily by shared symbols: the images of American flags that effectively divide the book into chapters, most of which are a little ragged and tattered (like America itself, they’ve seen better days) – and also by the cars and highways, by cowboys and jukeboxes. It's perhaps telling that the two photos of cowboys – both titled as connected to rodeo shows – were taken in Detroit and New York: that is, they're of men assuming the role as performers or audience members for a kind of travelling-circus version of the West.

As with Evans’s work, the images in The Americans are of public scenes. Unlike Evans’s, many of them depict interiors, but these are also public places: not houses, but cafes, bars, diners, restaurants, railway cars, hotels, municipal buildings. There are no domestic scenes or posed portraits at all. Cars are ubiquitous, partly because Frank’s project was conceived and executed as a series of road trips. There are plenty of cars in Evans too, but they were not so central to American life in the 1930s as they had become by 1955–6, when Frank worked. The paraphernalia of car culture is also everywhere: gas stations, rest-stop cafes, road intersections, casualties of road accidents. Again, Evans photographed car junkyards and garages, but it’s more insistent in Frank, and the infrastructure of automobile culture has become more complex. By contrast, a tram, a ferry, and a train interior all appear exactly once – and buses and aeroplanes are completely absent, though there are several photographs of urban pedestrians.

Unlike Evans, Frank shows little interest in architecture as a subject in and of itself – and despite the emphasis on cars, the suburbs are entirely absent. As the title of the book suggests, the majority of photographs have a human presence as their animating subject: that is, even if such subjects only occupy a small part of the frame, they are seemingly the justification for the photograph’s existence. There are several restricted milieux where, without the support of an institution like the FSA, Frank must have negotiated access himself (the 1956 Democratic convention in Chicago, a Detroit factory, a funeral among the black community). 

In The Photobook: A History, Gerry Badger places The Americans with Ed van der Elsken's Love on the Left Bank as an example of 'stream-of-consciousness' photography – indeed as the definitive example of this mode, which Badger links to Existentialism in philosophy, to Abstract Expressionism in painting, to bebop jazz in music, and to the Beat movement in literature. Frank's book is certainly 'rough, raw and gestural ... spontaneous and immediate, highly personal ... exciting, expressive, flying in the face of accepted photographic good taste' (233). But there's something here that goes against the stream-of-consciousness characterisation, as there also is in van der Elsken's book. In The Americans, although a minority of the images are of people glancing back at the camera, for the most part they’re not about the subjects’ response to Frank, and unlike in Love on the Left Bank, many images are quite clearly of people unaware of being photographed. Indeed, these rely on our sense of Frank’s invisibility for their force: of overlooking people caught off-guard, eating alone in a San Francisco cafeteria or sleeping in a park with their shoes off. Kerouac has it right: Frank 'sneaks around'; he has the 'strange secrecy of a shadow'. And this self-effacement seems at odds with the idea of an insistent authorial voice.

Some of Frank’s first viewers complained that this was a negative and ugly view of America: Papageorge quotes initial reviews that dismissed the book as 'warped', 'sick', 'neurotic' and 'joyless'. Some of the images are certainly ironic, if not satirical: besides the union bosses above, we might note the sailor at a Navy recruitment office visible only from his feet propped up on a desk ('Join the Navy Ask me about it' says the sticker on the side of the desk); or the racially segregated streetcar in New Orleans whose occupants all look separately at Frank, caged between the bars of its windows. This is another one of the images where Frank’s acknowledged presence is crucial to the effect, although it also relies upon his separateness from the subjects. But as with Evans, one senses that everything included here was of interest to Frank: he didn’t photograph out of enmity or disgust, but curiosity.

Evans used a variety of cameras, which required different kinds of engagement with his subjects. By contrast, every photograph in The Americans is taken from a 35mm negative, and, if I’m not mistaken, all of them were shot handheld and without flash. There's some debate online about what lenses Frank used: I thought (and to me the visual appearance of the photographs confirms this) that most of the images were shot on a ‘normal’ 50mm lens, one whose angle of view corresponds roughly to how we directly perceive a scene. But some images may have been taken on a wider 35mm lens, and a few are certainly telephoto compositions (e.g. a group at a cocktail party seemingly shot from a distance). 

In using available light and preferring a 50mm lens, Frank’s approach echoes that of Henri Cartier-Bresson, whose The Decisive Moment had been published in 1952 and was immediately promoted to classic status. But Frank was otherwise reacting against Cartier-Bresson, of whom he once said: 'you never felt he was moved by something that was happening other than the beauty of it, or just the composition.' And Frank was in general suspicious of a photographic practice rooted in photojournalism, 'those goddamned stories with a beginning and an end'. So he wasn’t looking for elegant compositions, and there’s little overt drama or conflict in The Americans. Instead of decisive moments, Frank depicts (in the words of Jno Cook) 'America stood still, frozen into a frightful pose between moments'. Papageorge expands on what this means in practice:

The characteristic gestures in his pictures are the slight, telling motions of the head and upper body: a glance (19, 37 [Papageorge refers here to page numbers in the first edition, though my Steidl reprint is unpaginated]), a stare (15, 41), a hand brought to the face (31, 51, 53), an arched neck (17, 55), pursed lips (15, 31). They suggest that Frank, like Evans, believed significance in a photograph might be consonant with the repose of the things it described. 

One of Cartier-Bresson’s standard techniques was to identify a promising location or background, and then wait there for something interesting to happen – for a dynamic element to appear, and to interact with and play off the carefully arranged background, as in the image below, from 1932.

There’s little sense of Frank doing this. Where his compositions do seem superficially similar to Cartier-Bresson’s, as in the image of the naval recruitment office, it feels more like genuine happenstance. Perhaps he was interested in the flag, and the shoes were either there already, or arrived serendipitously. It’s difficult to imagine him hovering around the doorway waiting, not least because while one may anticipate the arrival of a bicycle (or a car, or a pedestrian) on a street, one would never imagine the shoes in advance.

The French title of The Decisive Moment was Images à la Sauvette, on the fly, the French idiom referred to a street peddler without a license, like the bag-sellers in Venice. And that original title perhaps suits Frank’s book better than it does Cartier-Bresson’s, even if Frank had arranged access in advance to several of his locations – but then many of Cartier-Bresson’s images were borrowed from photojournalistic assignments, something that is more obvious from the selection in The Decisive Moment than in subsequent best-of anthologies.

There are other differences. Many of Cartier-Bresson’s images aspire to a quality of lightness or grace. Frank’s images are heavy. Even by the looser standards of photojournalism, let alone those of fine-art photography, the reproductions sometimes appear underexposed. Presumably shot on the fastest available films (probably with a speed of only ISO200), many were likely push-processed. Some have poor tonal separation and poor acutance (sharpness or edge definition). The textures of objects are sometimes visible – e.g. the silk (?) protective cover draped over a Long Beach car, and the wool suit and barbered nape of a man on a ferry – but they have to compete with the texture of the medium, i.e. visible, intrusive film grain.

Incidentally, this image is one of my favourites in The Americans, for the line of hats and the intimacy with which the boy is pressed in against what I take to be his father’s back, ensuring physical contact even as each of them are absorbed in their own thoughts. But again this is not a dramatic gesture – it's hardly a gesture at all.

Frank’s aesthetic follows inevitably from the decision to use only available light and shoot in poorly lit interiors. In other words, Frank insisted on making his photographs depict the things and places he was interested in, even under conditions where other photographers would just have given up and gone home. Muddy is the word that comes to mind: not only for the tones, but for the sense that the world is full of inertia and resistance, so that existence is something you have to wade through and struggle against, and not just with a camera. 

If I have reservations about Frank's approach, it’s with regard to the images without people. The rough tonalities, dead shadows and blown highlights of the small negative seem less justified when there’s no ephemeral human element that demands to be captured on the fly. But presumably for Frank consistency was more important than Evans’s variety. Don’t muck about: master one approach, one camera, one lens.

Like American Photographs, The Americans has one image per spread, so that each photograph is first seen in isolation as an independent visual statement before one is tempted to consider its relation to others in the sequence. And the titles are similarly minimalist, although here they are displayed opposite the images instead of separately at the end. The original French publication of the book juxtaposed the images with texts excerpted from a variety of commentators on America from various epochs, sourced and chosen by an editor, but thankfully these were removed for the American edition. Frank’s position has always been that a photograph should ‘nullify explanation’: that is, its effect should be primarily visual, untranslatable into language. The titles are therefore often limited to a bare statement of geography; sometimes they also specify the type of location or a social context ('Cocktail party – New York City', 'View from hotel window – Butte, Montana'), but they almost never attempt to describe or label the individuals depicted, so we are left to make our own inferences about the relationships between them. (An exception to this is 'Jehovah's Witness – Los Angeles', but since the man is visibly engaged in distributing copies of the movement's newspaper Awake, he's effectively labelling himself. Also the image on a ferry shown above is titled 'Yom Kippur – East River, New York City', which in effect identifies its subjects as Jewish, as otherwise it would not be relevant that it was taken on the Day of Atonement.)

The Americans is certainly arranged as a sequence. There are many pointed transitions: e.g. from an older man standing alone under the stairs of a Los Angeles rooming house to another on a bench at Yale wistfully watching a line of new students in black filing past; from a pair of hitchhikers driving Frank's car to another car zooming past in the background behind a bench of sedentary Florida pensioners, followed by another car draped with a protective silk cover, then a victim of a road accident draped with a blanket, which is in turn followed by a luminous shot of an empty highway receding into the distance. Or the kneeling supplicant holding a rudimentary cross by the Mississippi, which is then followed by a silhouetted statue of St Francis holding a cross up to bless a Los Angeles road intersection, and then a shaft of sunlight illuminating three crosses marking the site of a road accident, followed by an interior of a Detroit car-factory assembly line. As Badger explains, the flow is looser and more recursive than it is in American Photographs, frequently doubling back on itself:

Ideas ebb and flow, are introduced, discarded, recapitulated, transfigured, transposed, played off and piled up against each other with the exuberant energy and precise articulation of a Charlie Parker saxophone solo. (247)

Perhaps the larger rhythm of recurrent motifs and themes, with the book split into 'chapters', each introduced with an image of the American flag, is overall more important than specific transitions from one image to the next. We might say the same of much of American Photographs, but Frank's sequence retains both its variety and its coherence throughout, without succumbing to the repetitiveness that arguably undermines the latter part of Evans’s work. 

We could perhaps describe Evans as Apollonian – characterised by sobriety and lucid description – and Frank as Dionysian – characterised by passionate intensity. Certainly the introduction by Kerouac fits this description, and also seems well-suited to the visual aesthetic. But I don't see The Americans as a work of intoxication. Frank’s response to the scenes he photographs is his alone – with very few exceptions, he doesn't share himself with anyone he depicts. So if there's a loss of self involved it is, as with van der Elsken, in the form of a kind of erasure rather than an ecstatic communion with the world(s) depicted. But unlike in the work of Brassaï and van der Elsken, none of these images are collaborations with the subjects – nor do they seem directed or stage-managed (with the possible exception of the image of the three trans people on both counts). And whereas for the earlier photographers, subcultures constitute a genuine alternative to the mainstream, a parallel and independent mode of existence, many of Frank’s subjects seem more marginal. Partly this is because Brassaï and van der Elsken photograph their worlds from close up, whereas Frank catches them in passing.

Evans found it difficult, if not impossible, to match the achievement of American Photographs. Frank didn’t even try to match The Americans, shifting his artistic aspirations to filmmaking shortly after the book's publication. Though perhaps this usual summation of his career is too simple: he continued to work as a photographer for hire, or so I infer from a famous anecdote by Joel Meyerowitz, whose own career as a photographer was inspired by watching Frank do a shoot for the ad agency where Meyerowitz worked in 1962. Indeed there's plenty of evidence online showing him doing commercial work much later than this. But he didn't publish this work in books.

Five Wounds: Heraldry

In a previous post, I showed some of the sketches for the heraldic shields in Five Wounds. To create these shields, Dan and I had to learn the visual code of heraldry, which is governed by strict rules about how its various elements may be combined. These rules are analogous to a grammar. Indeed, heraldry is one of the few instances where this analogy really works when applied to a visual code. Heraldry is therefore a highly distinctive semiotic system, one might even say a uniquely pure semiotic system, and as such it represents one of the earliest historical examples of a coherent system of graphic design. 

According to the classic analysis of Ferdinand de Saussure, the signs that make up any given semiotic system can be broken down into two elements: the signifier (the actual sign, e.g. the word ‘dog’), and the signified (the concept that the sign represents, e.g. the dictionary definition of the word ‘dog’). A third possible element is the referent (the thing to which the concept refers, e.g. an actual dog), but for Saussure, language actually makes more sense if you think of it purely in terms of the relationship between signifier and signified, and exclude consideration of the referent altogether. 

What makes heraldry ‘uniquely pure’ is that it anticipates this conclusion. Nothing in heraldry claims to represent anything external to the code itself, or at least the referential aspect of the system is attenuated, as if heraldry has already evolved beyond this primitive function, which remains only in a vestigial form, like the tail on a human skeleton. 

How, then, does heraldry work? It is best understood as a code, like a computer code, which is used to generate shields or coats-of-arms. The sine qua non of heraldry is therefore the shield itself, but the most basic elements of the code, which are combined to create the shield, are the heraldic tinctures, which are split into two basic groups: colours and metals. The colours are blue (azure), red (gules), black (sable), and green (vert), although the last is used far less frequently than the first three. Some versions also add purple (purpure) to the colours, but it almost never occurs in actual historical shields. The metals are silver (argent) and gold (or), which may also be represented by white and yellow respectively. 

Every shield has one of these tinctures as its base or field, onto which are laid successive layers of additional elements – like layers in Photoshop – all of which are also assigned a tincture, and which cumulatively make up the shield. Thus the shield is always viewed as a single, combined image, but its component elements can always be broken down into a series of two-dimensional layers laid on top of one another in a predetermined order. From top to bottom, this order runs thus: field, ordinaries, subordinaries, charges. 

Ordinaries consist of geometric devices such as a bend (a diagonal stripe across the shield), a pale (a vertical stripe down the middle of the shield), a fess (a horizontal strip across the middle of the shield), a cross, a saltire (a diagonal cross), a chevron, and so on. These are all laid on top of the field, and one of their functions is to subdivide it, so that their proportions in relation to the total area of the shield are therefore strictly controlled. For example, a pale or a fess should occupy roughly one-third of the shield’s area, while a chevron should occupy roughly one-fifth. Subordinaries are also geometric devices. They include the chief (a horizontal strip across the top third of the shield), the canton (a small square in the dexter chief, i.e. the top left from the viewer’s perspective [1]) and the bordure (a thin border around the shield’s outer edge). 

The final layer is that of pictorial charges. The most common of these are lions and eagles. Charges do have a notional referent, but they are always rendered in a highly abstract manner, and their relation to actual lions and eagles, or even to the symbolic meanings conventionally associated with lions and eagles in medieval bestiaries is, in effect, ‘bracketed’: it is irrelevant for how the pictorial code actually functions. 

The field is not always an undifferentiated, single tincture. It can also be divided in various conventional ways. For example, a shield divided ‘per bend’ has its field split in two diagonally, whereas a shield divided ‘per pale’ has its field split in two vertically. 

The ‘rule of tincture’ governs the way in which tinctures may be assigned to the various layers of a shield. It states that a colour cannot be laid directly on top of another colour; nor can a metal be laid directly on top of a metal. So if the field is a metal, any ordinary laid on top of it must be a colour, while any charge laid on top of that must in turn be a metal. There are, however, some exceptions to this rule, notably that it does not apply to subordinaries, or to divisions of the field. [2]  

In heraldry, blue represents only the idea of blue, and red the idea of red, and it does not matter which specific shade of blue or red is used to embody that idea. Indeed, in the seventeenth century, when it became common practice to print compilations of coats-of-arms in black-and-white reproductions, a system of cross-hatchings was invented to represent the tinctures, and these cross-hatchings represent their respective associated pigments perfectly: that is, in heraldry, a regularly-distributed pattern of dots signifies the idea of gold just as adequately as any particular yellow or gold pigment can do. 

As I have described it thus far, heraldry is a purely visual system, but – to return to our metaphor of computer code – it has a written analogue, whereby a description of a shield can be created, which serves as a programme, whose output is the visual representation of the shield.

So, for example, the shield above is encoded as ‘gules, a bordure argent’, which means ‘a red field with a silver border’; whereas the shield below is ‘Per pale sable and argent, on a chief vert a canton or’, which means ‘a field divided in two vertically, with black on the left and and silver on the right, which is surmounted by a green strip occupying the top third, which is in turn surmounted by a gold square in the top left corner’.

Any written description should allow you – if it is parsed correctly – to generate the output of a correct shield with total accuracy. And as with computer code, any ambiguity in the initial command constitutes a fatal error, while any elements or aspects of the output that are not predetermined by this code are by definition irrelevant. Minute variations in the written code may have larger consequences when it is translated into a shield. For example, the first shield below is ‘argent, a bend sable’; conversely, the second is ‘Per bend, argent and sable’; the third is ‘Per bendy, sable and argent’; while the last is ‘Per bendy, argent and sable’.

According to Shaun Tan, an image only works insofar as it can't be reduced to a written description (an argument that recalls Robert Frank's insistence that a photograph should nullify explanation), but heraldry is based upon the opposite assumption. As such, it offers one possible way of exploring the relationship between word and image, which is both a recurrent theme and a practical challenge in Five Wounds.

So why did Dan and I use heraldry in Five Wounds, and to what ends? 

When I was looking for a way to represent the book’s structure visually, I needed a system with seven elements, but that in practice only made use of five them regularly: because there are five protagonists in my book, but two other characters whose status is sometimes brought into question. Moreover, of my central five characters, I needed to be able to subdivide them into two groups, of three and two respectively, because this is how their various relationships play out. 

All of this found an analogue in the heraldic code, in which there are seven basic tinctures, but only five that occur with any regularity, and of those five, there are three colours and two metals. 

This sounds suspiciously convenient. It is possible that, at an unconscious level, I was already thinking of how to relate the structure of the book to that of heraldry right from the start, so the correspondence may not be entirely coincidental. 

In any case, thinking that the heraldic code could perhaps be converted into a map of the book’s structure, I began to develop this idea by assigning a tincture to each of my protagonists: blue to the mutilated angel, Gabriella; red to the man of blood, Cur; black to the amorphous man with a wax face, Cuckoo; silver to the daguerreotypist, Magpie; and gold to the alchemist, Crow. These designations loosely follow traditional colour symbolism, so they have a logic of sorts, although such symbolism is not part of the heraldic code. 

My next step was to design five shields, one for each of the protagonists, each using the schema outlined above: that is, Gabriella’s shield has a blue field, while Cur’s has a red field, and so on; and the addition of other elements (ordinaries, subordinaries and charges) onto the field in other tinctures would then, in each case, map the various relationships between the five protagonists.

  

Above: Gabriella’s and Magpie’s motto shields

Along with these ‘motto shields’ (designated thus by Dan and I for obvious reasons), there is an entirely separate series of smaller shields, which open each chapter (in the first edition of the novel, there were more of these: one for each layout). These serve as a visual index of which characters appear in the chapter in question.

There are important differences between the motto shields and the index shields. The former use pictorial charges to help define the protagonists: wings for the angel, Gabriella; a wolf for the rabid Cur; the moon and stars for the nocturnal daguerreotypist, Magpie; and so on. By contrast, the index shields are purely abstract, consisting only of the field, ordinaries and subordinaries. And, whereas the motto shields use only five tinctures to refer to five protagonists, the index shields use all seven tinctures, which therefore refer to seven different characters. 

There are therefore two overlapping but separate indexical colour systems in Five Wounds: one with five elements, and one with seven elements. 

Every index shield is unique. When different chapters involve the same group of characters, and thus use the same tinctures, each of these is represented by a different shield design. And even for the chapters in which only one character appears, and which are therefore indexed with an undifferentiated shield of a single tincture, I had Dan repaint the shield every time we used it, so that the patterning of the pigment would be slightly different. 

Having gone to the trouble of learning the language of heraldry, it may seem perverse that the first thing I decided to do with it was to violate its integrity by forcing it to describe something external to itself: that is, my five (or is it seven?) protagonists. In Five Wounds, then, the language of heraldry is persistently construed wrongly. Dan and I deliberately created ‘ungrammatical’ visual statements, which are the inevitable consequence of forcing the code to express things it was never intended to express. 

There are, for example, several index shields that violate the rule of tincture by combining colour with colour and metal with metal, depending on which characters happen to appear in any given chapter: that is, according to a criterion that is entirely irrelevant to the internal logic of the code. 

Forcing heraldry to perform such an unnatural function creates interesting problems, which were highlighted for me when I tried to encode the motto shields I designed for the five protagonists in written descriptions thereof. Below is the horrific result of this exercise, which illustrates the impossibility of trying to describe a series of (deliberate) mistakes using a code whose entire purpose is to eliminate ambiguity. It can’t be done.

  

Above: A failed attempt to provide written descriptions for the five motto shields

This unreadable attempt at definition is not included in Five Wounds, but it inspired one of several handwritten annotations added to the text of the first edition of the novel, in which a garbled representation of the five motto shields (drawn by Dan) is accompanied by the cryptic note, The problem with a perfect notation system: It can’t describe an error. Of course, this observation has much broader implications in the context of the novel than its application to heraldry.

  

Above: annotation from the end of the first edition of Five Wounds

However, I've cut this from the second edition, as, besides being beyond my powers to incorporate when I re-typeset the text, it also seemed a little too cute.

If a work is to be coherent, then certain ideas have to present in its DNA: they have to run through every aspect of its narrative and presentation. In Five Wounds, these ideas include: translation, garbled transmission, insecure attributions of meaning, the relationship between signal and noise, the nature of interpretation, and what constitutes a misinterpretation. It should be obvious how the book’s use – and abuse – of heraldry helps to dramatise some of these themes. 

If the colours represent the book’s protagonists, they also necessarily oversimplify their interactions by rendering them in diagrammatic form. Thus the colour coding is not only a misinterpretation because it makes improper use of heraldry: it is also a misinterpretation because, in doing so, it reduces each of the characters to a single, deterministic attribute. 

None of this would be of any interest, of course, unless the heraldic shields communicate something to the viewer emphatically and immediately, as images; but that is the advantage of using a visual code designed to do exactly that. 

Postscript: At some point, I'll provide a separate explanation of why Dan’s colour renditions of the shields are painted (deliberately) incompetently, with the paint spilling over the borders indicated by the underlying patterns. Also, see here for another short analysis of heraldry as a semiotic system.

[1] ‘Dexter’ actually designates the right-hand side of the shield, and ‘sinister’ the left-hand side, but these are determined from the point-of-view of the fictional person holding the shield out in front of them towards the viewer. I have therefore simplified matters by sticking to the viewer's point-of-view here. This account of heraldry is also simplified in other respects.  

[2] It does not apply to divisions of the field in theory. In practice, I have never seen a divided field that does not alternate a metal and a colour for the division.